2009-04-28 09:57:12 +07:00
|
|
|
|
/*
|
|
|
|
|
* xHCI host controller driver
|
|
|
|
|
*
|
|
|
|
|
* Copyright (C) 2008 Intel Corp.
|
|
|
|
|
*
|
|
|
|
|
* Author: Sarah Sharp
|
|
|
|
|
* Some code borrowed from the Linux EHCI driver.
|
|
|
|
|
*
|
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
|
* it under the terms of the GNU General Public License version 2 as
|
|
|
|
|
* published by the Free Software Foundation.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful, but
|
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
|
|
|
|
|
* or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
|
|
|
* for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
|
* along with this program; if not, write to the Free Software Foundation,
|
|
|
|
|
* Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|
|
|
|
*/
|
|
|
|
|
|
2014-05-08 23:26:00 +07:00
|
|
|
|
|
|
|
|
|
#include <linux/slab.h>
|
2009-04-28 09:57:12 +07:00
|
|
|
|
#include <asm/unaligned.h>
|
|
|
|
|
|
|
|
|
|
#include "xhci.h"
|
2013-08-06 11:52:45 +07:00
|
|
|
|
#include "xhci-trace.h"
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
2010-10-14 21:23:03 +07:00
|
|
|
|
#define PORT_WAKE_BITS (PORT_WKOC_E | PORT_WKDISC_E | PORT_WKCONN_E)
|
|
|
|
|
#define PORT_RWC_BITS (PORT_CSC | PORT_PEC | PORT_WRC | PORT_OCC | \
|
|
|
|
|
PORT_RC | PORT_PLC | PORT_PE)
|
|
|
|
|
|
2012-08-22 20:12:06 +07:00
|
|
|
|
/* USB 3.0 BOS descriptor and a capability descriptor, combined */
|
2011-10-07 01:54:23 +07:00
|
|
|
|
static u8 usb_bos_descriptor [] = {
|
|
|
|
|
USB_DT_BOS_SIZE, /* __u8 bLength, 5 bytes */
|
|
|
|
|
USB_DT_BOS, /* __u8 bDescriptorType */
|
|
|
|
|
0x0F, 0x00, /* __le16 wTotalLength, 15 bytes */
|
|
|
|
|
0x1, /* __u8 bNumDeviceCaps */
|
|
|
|
|
/* First device capability */
|
|
|
|
|
USB_DT_USB_SS_CAP_SIZE, /* __u8 bLength, 10 bytes */
|
|
|
|
|
USB_DT_DEVICE_CAPABILITY, /* Device Capability */
|
|
|
|
|
USB_SS_CAP_TYPE, /* bDevCapabilityType, SUPERSPEED_USB */
|
|
|
|
|
0x00, /* bmAttributes, LTM off by default */
|
|
|
|
|
USB_5GBPS_OPERATION, 0x00, /* wSpeedsSupported, 5Gbps only */
|
|
|
|
|
0x03, /* bFunctionalitySupport,
|
|
|
|
|
USB 3.0 speed only */
|
|
|
|
|
0x00, /* bU1DevExitLat, set later. */
|
|
|
|
|
0x00, 0x00 /* __le16 bU2DevExitLat, set later. */
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
2010-11-30 07:14:37 +07:00
|
|
|
|
static void xhci_common_hub_descriptor(struct xhci_hcd *xhci,
|
|
|
|
|
struct usb_hub_descriptor *desc, int ports)
|
2009-04-28 09:57:12 +07:00
|
|
|
|
{
|
|
|
|
|
u16 temp;
|
|
|
|
|
|
|
|
|
|
desc->bPwrOn2PwrGood = 10; /* xhci section 5.4.9 says 20ms max */
|
|
|
|
|
desc->bHubContrCurrent = 0;
|
|
|
|
|
|
|
|
|
|
desc->bNbrPorts = ports;
|
|
|
|
|
temp = 0;
|
2011-11-22 21:03:36 +07:00
|
|
|
|
/* Bits 1:0 - support per-port power switching, or power always on */
|
2009-04-28 09:57:12 +07:00
|
|
|
|
if (HCC_PPC(xhci->hcc_params))
|
2011-11-22 21:03:36 +07:00
|
|
|
|
temp |= HUB_CHAR_INDV_PORT_LPSM;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
else
|
2011-11-22 21:03:36 +07:00
|
|
|
|
temp |= HUB_CHAR_NO_LPSM;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
/* Bit 2 - root hubs are not part of a compound device */
|
|
|
|
|
/* Bits 4:3 - individual port over current protection */
|
2011-11-22 21:03:36 +07:00
|
|
|
|
temp |= HUB_CHAR_INDV_PORT_OCPM;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
/* Bits 6:5 - no TTs in root ports */
|
|
|
|
|
/* Bit 7 - no port indicators */
|
2011-03-29 09:40:46 +07:00
|
|
|
|
desc->wHubCharacteristics = cpu_to_le16(temp);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
}
|
|
|
|
|
|
2010-11-30 07:14:37 +07:00
|
|
|
|
/* Fill in the USB 2.0 roothub descriptor */
|
|
|
|
|
static void xhci_usb2_hub_descriptor(struct usb_hcd *hcd, struct xhci_hcd *xhci,
|
|
|
|
|
struct usb_hub_descriptor *desc)
|
|
|
|
|
{
|
|
|
|
|
int ports;
|
|
|
|
|
u16 temp;
|
|
|
|
|
__u8 port_removable[(USB_MAXCHILDREN + 1 + 7) / 8];
|
|
|
|
|
u32 portsc;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
ports = xhci->num_usb2_ports;
|
|
|
|
|
|
|
|
|
|
xhci_common_hub_descriptor(xhci, desc, ports);
|
2011-11-22 21:03:36 +07:00
|
|
|
|
desc->bDescriptorType = USB_DT_HUB;
|
2010-11-30 07:14:37 +07:00
|
|
|
|
temp = 1 + (ports / 8);
|
2011-11-22 21:03:36 +07:00
|
|
|
|
desc->bDescLength = USB_DT_HUB_NONVAR_SIZE + 2 * temp;
|
2010-11-30 07:14:37 +07:00
|
|
|
|
|
|
|
|
|
/* The Device Removable bits are reported on a byte granularity.
|
|
|
|
|
* If the port doesn't exist within that byte, the bit is set to 0.
|
|
|
|
|
*/
|
|
|
|
|
memset(port_removable, 0, sizeof(port_removable));
|
|
|
|
|
for (i = 0; i < ports; i++) {
|
2013-11-15 10:34:06 +07:00
|
|
|
|
portsc = readl(xhci->usb2_ports[i]);
|
2010-11-30 07:14:37 +07:00
|
|
|
|
/* If a device is removable, PORTSC reports a 0, same as in the
|
|
|
|
|
* hub descriptor DeviceRemovable bits.
|
|
|
|
|
*/
|
|
|
|
|
if (portsc & PORT_DEV_REMOVE)
|
|
|
|
|
/* This math is hairy because bit 0 of DeviceRemovable
|
|
|
|
|
* is reserved, and bit 1 is for port 1, etc.
|
|
|
|
|
*/
|
|
|
|
|
port_removable[(i + 1) / 8] |= 1 << ((i + 1) % 8);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* ch11.h defines a hub descriptor that has room for USB_MAXCHILDREN
|
|
|
|
|
* ports on it. The USB 2.0 specification says that there are two
|
|
|
|
|
* variable length fields at the end of the hub descriptor:
|
|
|
|
|
* DeviceRemovable and PortPwrCtrlMask. But since we can have less than
|
|
|
|
|
* USB_MAXCHILDREN ports, we may need to use the DeviceRemovable array
|
|
|
|
|
* to set PortPwrCtrlMask bits. PortPwrCtrlMask must always be set to
|
|
|
|
|
* 0xFF, so we initialize the both arrays (DeviceRemovable and
|
|
|
|
|
* PortPwrCtrlMask) to 0xFF. Then we set the DeviceRemovable for each
|
|
|
|
|
* set of ports that actually exist.
|
|
|
|
|
*/
|
|
|
|
|
memset(desc->u.hs.DeviceRemovable, 0xff,
|
|
|
|
|
sizeof(desc->u.hs.DeviceRemovable));
|
|
|
|
|
memset(desc->u.hs.PortPwrCtrlMask, 0xff,
|
|
|
|
|
sizeof(desc->u.hs.PortPwrCtrlMask));
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < (ports + 1 + 7) / 8; i++)
|
|
|
|
|
memset(&desc->u.hs.DeviceRemovable[i], port_removable[i],
|
|
|
|
|
sizeof(__u8));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Fill in the USB 3.0 roothub descriptor */
|
|
|
|
|
static void xhci_usb3_hub_descriptor(struct usb_hcd *hcd, struct xhci_hcd *xhci,
|
|
|
|
|
struct usb_hub_descriptor *desc)
|
|
|
|
|
{
|
|
|
|
|
int ports;
|
|
|
|
|
u16 port_removable;
|
|
|
|
|
u32 portsc;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
ports = xhci->num_usb3_ports;
|
|
|
|
|
xhci_common_hub_descriptor(xhci, desc, ports);
|
2011-11-22 21:03:36 +07:00
|
|
|
|
desc->bDescriptorType = USB_DT_SS_HUB;
|
|
|
|
|
desc->bDescLength = USB_DT_SS_HUB_SIZE;
|
2010-11-30 07:14:37 +07:00
|
|
|
|
|
|
|
|
|
/* header decode latency should be zero for roothubs,
|
|
|
|
|
* see section 4.23.5.2.
|
|
|
|
|
*/
|
|
|
|
|
desc->u.ss.bHubHdrDecLat = 0;
|
|
|
|
|
desc->u.ss.wHubDelay = 0;
|
|
|
|
|
|
|
|
|
|
port_removable = 0;
|
|
|
|
|
/* bit 0 is reserved, bit 1 is for port 1, etc. */
|
|
|
|
|
for (i = 0; i < ports; i++) {
|
2013-11-15 10:34:06 +07:00
|
|
|
|
portsc = readl(xhci->usb3_ports[i]);
|
2010-11-30 07:14:37 +07:00
|
|
|
|
if (portsc & PORT_DEV_REMOVE)
|
|
|
|
|
port_removable |= 1 << (i + 1);
|
|
|
|
|
}
|
2012-10-15 14:38:35 +07:00
|
|
|
|
|
|
|
|
|
desc->u.ss.DeviceRemovable = cpu_to_le16(port_removable);
|
2010-11-30 07:14:37 +07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void xhci_hub_descriptor(struct usb_hcd *hcd, struct xhci_hcd *xhci,
|
|
|
|
|
struct usb_hub_descriptor *desc)
|
|
|
|
|
{
|
|
|
|
|
|
|
|
|
|
if (hcd->speed == HCD_USB3)
|
|
|
|
|
xhci_usb3_hub_descriptor(hcd, xhci, desc);
|
|
|
|
|
else
|
|
|
|
|
xhci_usb2_hub_descriptor(hcd, xhci, desc);
|
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
2009-04-28 09:57:12 +07:00
|
|
|
|
static unsigned int xhci_port_speed(unsigned int port_status)
|
|
|
|
|
{
|
|
|
|
|
if (DEV_LOWSPEED(port_status))
|
2010-03-04 23:32:30 +07:00
|
|
|
|
return USB_PORT_STAT_LOW_SPEED;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
if (DEV_HIGHSPEED(port_status))
|
2010-03-04 23:32:30 +07:00
|
|
|
|
return USB_PORT_STAT_HIGH_SPEED;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
/*
|
|
|
|
|
* FIXME: Yes, we should check for full speed, but the core uses that as
|
|
|
|
|
* a default in portspeed() in usb/core/hub.c (which is the only place
|
2010-03-04 23:32:30 +07:00
|
|
|
|
* USB_PORT_STAT_*_SPEED is used).
|
2009-04-28 09:57:12 +07:00
|
|
|
|
*/
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* These bits are Read Only (RO) and should be saved and written to the
|
|
|
|
|
* registers: 0, 3, 10:13, 30
|
|
|
|
|
* connect status, over-current status, port speed, and device removable.
|
|
|
|
|
* connect status and port speed are also sticky - meaning they're in
|
|
|
|
|
* the AUX well and they aren't changed by a hot, warm, or cold reset.
|
|
|
|
|
*/
|
|
|
|
|
#define XHCI_PORT_RO ((1<<0) | (1<<3) | (0xf<<10) | (1<<30))
|
|
|
|
|
/*
|
|
|
|
|
* These bits are RW; writing a 0 clears the bit, writing a 1 sets the bit:
|
|
|
|
|
* bits 5:8, 9, 14:15, 25:27
|
|
|
|
|
* link state, port power, port indicator state, "wake on" enable state
|
|
|
|
|
*/
|
|
|
|
|
#define XHCI_PORT_RWS ((0xf<<5) | (1<<9) | (0x3<<14) | (0x7<<25))
|
|
|
|
|
/*
|
|
|
|
|
* These bits are RW; writing a 1 sets the bit, writing a 0 has no effect:
|
|
|
|
|
* bit 4 (port reset)
|
|
|
|
|
*/
|
|
|
|
|
#define XHCI_PORT_RW1S ((1<<4))
|
|
|
|
|
/*
|
|
|
|
|
* These bits are RW; writing a 1 clears the bit, writing a 0 has no effect:
|
|
|
|
|
* bits 1, 17, 18, 19, 20, 21, 22, 23
|
|
|
|
|
* port enable/disable, and
|
|
|
|
|
* change bits: connect, PED, warm port reset changed (reserved zero for USB 2.0 ports),
|
|
|
|
|
* over-current, reset, link state, and L1 change
|
|
|
|
|
*/
|
|
|
|
|
#define XHCI_PORT_RW1CS ((1<<1) | (0x7f<<17))
|
|
|
|
|
/*
|
|
|
|
|
* Bit 16 is RW, and writing a '1' to it causes the link state control to be
|
|
|
|
|
* latched in
|
|
|
|
|
*/
|
|
|
|
|
#define XHCI_PORT_RW ((1<<16))
|
|
|
|
|
/*
|
|
|
|
|
* These bits are Reserved Zero (RsvdZ) and zero should be written to them:
|
|
|
|
|
* bits 2, 24, 28:31
|
|
|
|
|
*/
|
|
|
|
|
#define XHCI_PORT_RZ ((1<<2) | (1<<24) | (0xf<<28))
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Given a port state, this function returns a value that would result in the
|
|
|
|
|
* port being in the same state, if the value was written to the port status
|
|
|
|
|
* control register.
|
|
|
|
|
* Save Read Only (RO) bits and save read/write bits where
|
|
|
|
|
* writing a 0 clears the bit and writing a 1 sets the bit (RWS).
|
|
|
|
|
* For all other types (RW1S, RW1CS, RW, and RZ), writing a '0' has no effect.
|
|
|
|
|
*/
|
2010-10-14 21:23:00 +07:00
|
|
|
|
u32 xhci_port_state_to_neutral(u32 state)
|
2009-04-28 09:57:12 +07:00
|
|
|
|
{
|
|
|
|
|
/* Save read-only status and port state */
|
|
|
|
|
return (state & XHCI_PORT_RO) | (state & XHCI_PORT_RWS);
|
|
|
|
|
}
|
|
|
|
|
|
2010-10-14 21:22:57 +07:00
|
|
|
|
/*
|
|
|
|
|
* find slot id based on port number.
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
* @port: The one-based port number from one of the two split roothubs.
|
2010-10-14 21:22:57 +07:00
|
|
|
|
*/
|
2010-12-17 01:49:09 +07:00
|
|
|
|
int xhci_find_slot_id_by_port(struct usb_hcd *hcd, struct xhci_hcd *xhci,
|
|
|
|
|
u16 port)
|
2010-10-14 21:22:57 +07:00
|
|
|
|
{
|
|
|
|
|
int slot_id;
|
|
|
|
|
int i;
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
enum usb_device_speed speed;
|
2010-10-14 21:22:57 +07:00
|
|
|
|
|
|
|
|
|
slot_id = 0;
|
|
|
|
|
for (i = 0; i < MAX_HC_SLOTS; i++) {
|
|
|
|
|
if (!xhci->devs[i])
|
|
|
|
|
continue;
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
speed = xhci->devs[i]->udev->speed;
|
|
|
|
|
if (((speed == USB_SPEED_SUPER) == (hcd->speed == HCD_USB3))
|
2011-09-03 01:05:41 +07:00
|
|
|
|
&& xhci->devs[i]->fake_port == port) {
|
2010-10-14 21:22:57 +07:00
|
|
|
|
slot_id = i;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return slot_id;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Stop device
|
|
|
|
|
* It issues stop endpoint command for EP 0 to 30. And wait the last command
|
|
|
|
|
* to complete.
|
|
|
|
|
* suspend will set to 1, if suspend bit need to set in command.
|
|
|
|
|
*/
|
|
|
|
|
static int xhci_stop_device(struct xhci_hcd *xhci, int slot_id, int suspend)
|
|
|
|
|
{
|
|
|
|
|
struct xhci_virt_device *virt_dev;
|
|
|
|
|
struct xhci_command *cmd;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int ret;
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
|
virt_dev = xhci->devs[slot_id];
|
|
|
|
|
cmd = xhci_alloc_command(xhci, false, true, GFP_NOIO);
|
|
|
|
|
if (!cmd) {
|
|
|
|
|
xhci_dbg(xhci, "Couldn't allocate command structure.\n");
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
for (i = LAST_EP_INDEX; i > 0; i--) {
|
2014-05-08 23:26:00 +07:00
|
|
|
|
if (virt_dev->eps[i].ring && virt_dev->eps[i].ring->dequeue) {
|
|
|
|
|
struct xhci_command *command;
|
|
|
|
|
command = xhci_alloc_command(xhci, false, false,
|
2014-06-10 15:27:41 +07:00
|
|
|
|
GFP_NOWAIT);
|
2014-05-08 23:26:00 +07:00
|
|
|
|
if (!command) {
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
xhci_free_command(xhci, cmd);
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
xhci_queue_stop_endpoint(xhci, command, slot_id, i,
|
|
|
|
|
suspend);
|
|
|
|
|
}
|
2010-10-14 21:22:57 +07:00
|
|
|
|
}
|
2014-05-08 23:26:00 +07:00
|
|
|
|
xhci_queue_stop_endpoint(xhci, cmd, slot_id, 0, suspend);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
xhci_ring_cmd_db(xhci);
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
|
|
|
|
|
/* Wait for last stop endpoint command to finish */
|
xhci: rework command timeout and cancellation,
Use one timer to control command timeout.
start/kick the timer every time a command is completed and a
new command is waiting, or a new command is added to a empty list.
If the timer runs out, then tag the current command as "aborted", and
start the xhci command abortion process.
Previously each function that submitted a command had its own timer.
If that command timed out, a new command structure for the
command was created and it was put on a cancel_cmd_list list,
then a pci write to abort the command ring was issued.
when the ring was aborted, it checked if the current command
was the one to be canceled, later when the ring was stopped the
driver got ownership of the TRBs in the command ring,
compared then to the TRBs in the cancel_cmd_list,
and turned them into No-ops.
Now, instead, at timeout we tag the status of the command in the
command queue to be aborted, and start the ring abortion.
Ring abortion stops the command ring and gives control of the
commands to us.
All the aborted commands are now turned into No-ops.
If the ring is already stopped when the command times outs its not possible
to start the ring abortion, in this case the command is turnd to No-op
right away.
All these changes allows us to remove the entire cancel_cmd_list code.
The functions waiting for a command to finish no longer have their own timeouts.
They will wait either until the command completes normally,
or until the whole command abortion is done.
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-05-08 23:26:03 +07:00
|
|
|
|
wait_for_completion(cmd->completion);
|
|
|
|
|
|
|
|
|
|
if (cmd->status == COMP_CMD_ABORT || cmd->status == COMP_CMD_STOP) {
|
|
|
|
|
xhci_warn(xhci, "Timeout while waiting for stop endpoint command\n");
|
2010-10-14 21:22:57 +07:00
|
|
|
|
ret = -ETIME;
|
|
|
|
|
}
|
|
|
|
|
xhci_free_command(xhci, cmd);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Ring device, it rings the all doorbells unconditionally.
|
|
|
|
|
*/
|
2010-10-14 21:23:00 +07:00
|
|
|
|
void xhci_ring_device(struct xhci_hcd *xhci, int slot_id)
|
2010-10-14 21:22:57 +07:00
|
|
|
|
{
|
2014-08-20 20:41:56 +07:00
|
|
|
|
int i, s;
|
|
|
|
|
struct xhci_virt_ep *ep;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < LAST_EP_INDEX + 1; i++) {
|
|
|
|
|
ep = &xhci->devs[slot_id]->eps[i];
|
2010-10-14 21:22:57 +07:00
|
|
|
|
|
2014-08-20 20:41:56 +07:00
|
|
|
|
if (ep->ep_state & EP_HAS_STREAMS) {
|
|
|
|
|
for (s = 1; s < ep->stream_info->num_streams; s++)
|
|
|
|
|
xhci_ring_ep_doorbell(xhci, slot_id, i, s);
|
|
|
|
|
} else if (ep->ring && ep->ring->dequeue) {
|
2010-10-14 21:22:57 +07:00
|
|
|
|
xhci_ring_ep_doorbell(xhci, slot_id, i, 0);
|
2014-08-20 20:41:56 +07:00
|
|
|
|
}
|
|
|
|
|
}
|
2010-10-14 21:22:57 +07:00
|
|
|
|
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
static void xhci_disable_port(struct usb_hcd *hcd, struct xhci_hcd *xhci,
|
2011-03-29 09:40:46 +07:00
|
|
|
|
u16 wIndex, __le32 __iomem *addr, u32 port_status)
|
2009-12-10 06:59:11 +07:00
|
|
|
|
{
|
2010-11-17 06:58:52 +07:00
|
|
|
|
/* Don't allow the USB core to disable SuperSpeed ports. */
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
if (hcd->speed == HCD_USB3) {
|
2010-11-17 06:58:52 +07:00
|
|
|
|
xhci_dbg(xhci, "Ignoring request to disable "
|
|
|
|
|
"SuperSpeed port.\n");
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-10 06:59:11 +07:00
|
|
|
|
/* Write 1 to disable the port */
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(port_status | PORT_PE, addr);
|
2013-11-15 10:34:06 +07:00
|
|
|
|
port_status = readl(addr);
|
2009-12-10 06:59:11 +07:00
|
|
|
|
xhci_dbg(xhci, "disable port, actual port %d status = 0x%x\n",
|
|
|
|
|
wIndex, port_status);
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-10 06:59:08 +07:00
|
|
|
|
static void xhci_clear_port_change_bit(struct xhci_hcd *xhci, u16 wValue,
|
2011-03-29 09:40:46 +07:00
|
|
|
|
u16 wIndex, __le32 __iomem *addr, u32 port_status)
|
2009-12-10 06:59:08 +07:00
|
|
|
|
{
|
|
|
|
|
char *port_change_bit;
|
|
|
|
|
u32 status;
|
|
|
|
|
|
|
|
|
|
switch (wValue) {
|
|
|
|
|
case USB_PORT_FEAT_C_RESET:
|
|
|
|
|
status = PORT_RC;
|
|
|
|
|
port_change_bit = "reset";
|
|
|
|
|
break;
|
2011-04-27 17:07:29 +07:00
|
|
|
|
case USB_PORT_FEAT_C_BH_PORT_RESET:
|
|
|
|
|
status = PORT_WRC;
|
|
|
|
|
port_change_bit = "warm(BH) reset";
|
|
|
|
|
break;
|
2009-12-10 06:59:08 +07:00
|
|
|
|
case USB_PORT_FEAT_C_CONNECTION:
|
|
|
|
|
status = PORT_CSC;
|
|
|
|
|
port_change_bit = "connect";
|
|
|
|
|
break;
|
|
|
|
|
case USB_PORT_FEAT_C_OVER_CURRENT:
|
|
|
|
|
status = PORT_OCC;
|
|
|
|
|
port_change_bit = "over-current";
|
|
|
|
|
break;
|
2009-12-10 06:59:11 +07:00
|
|
|
|
case USB_PORT_FEAT_C_ENABLE:
|
|
|
|
|
status = PORT_PEC;
|
|
|
|
|
port_change_bit = "enable/disable";
|
|
|
|
|
break;
|
2010-10-14 21:22:57 +07:00
|
|
|
|
case USB_PORT_FEAT_C_SUSPEND:
|
|
|
|
|
status = PORT_PLC;
|
|
|
|
|
port_change_bit = "suspend/resume";
|
|
|
|
|
break;
|
2011-04-27 17:07:35 +07:00
|
|
|
|
case USB_PORT_FEAT_C_PORT_LINK_STATE:
|
|
|
|
|
status = PORT_PLC;
|
|
|
|
|
port_change_bit = "link state";
|
|
|
|
|
break;
|
2015-03-23 23:27:41 +07:00
|
|
|
|
case USB_PORT_FEAT_C_PORT_CONFIG_ERROR:
|
|
|
|
|
status = PORT_CEC;
|
|
|
|
|
port_change_bit = "config error";
|
|
|
|
|
break;
|
2009-12-10 06:59:08 +07:00
|
|
|
|
default:
|
|
|
|
|
/* Should never happen */
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
/* Change bits are all write 1 to clear */
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(port_status | status, addr);
|
2013-11-15 10:34:06 +07:00
|
|
|
|
port_status = readl(addr);
|
2009-12-10 06:59:08 +07:00
|
|
|
|
xhci_dbg(xhci, "clear port %s change, actual port %d status = 0x%x\n",
|
|
|
|
|
port_change_bit, wIndex, port_status);
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-03 20:11:00 +07:00
|
|
|
|
static int xhci_get_ports(struct usb_hcd *hcd, __le32 __iomem ***port_array)
|
|
|
|
|
{
|
|
|
|
|
int max_ports;
|
|
|
|
|
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
|
|
|
|
|
|
|
|
|
|
if (hcd->speed == HCD_USB3) {
|
|
|
|
|
max_ports = xhci->num_usb3_ports;
|
|
|
|
|
*port_array = xhci->usb3_ports;
|
|
|
|
|
} else {
|
|
|
|
|
max_ports = xhci->num_usb2_ports;
|
|
|
|
|
*port_array = xhci->usb2_ports;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return max_ports;
|
|
|
|
|
}
|
|
|
|
|
|
2011-09-24 04:19:48 +07:00
|
|
|
|
void xhci_set_link_state(struct xhci_hcd *xhci, __le32 __iomem **port_array,
|
|
|
|
|
int port_id, u32 link_state)
|
|
|
|
|
{
|
|
|
|
|
u32 temp;
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[port_id]);
|
2011-09-24 04:19:48 +07:00
|
|
|
|
temp = xhci_port_state_to_neutral(temp);
|
|
|
|
|
temp &= ~PORT_PLS_MASK;
|
|
|
|
|
temp |= PORT_LINK_STROBE | link_state;
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[port_id]);
|
2011-09-24 04:19:48 +07:00
|
|
|
|
}
|
|
|
|
|
|
2012-08-07 18:10:03 +07:00
|
|
|
|
static void xhci_set_remote_wake_mask(struct xhci_hcd *xhci,
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
__le32 __iomem **port_array, int port_id, u16 wake_mask)
|
|
|
|
|
{
|
|
|
|
|
u32 temp;
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[port_id]);
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
temp = xhci_port_state_to_neutral(temp);
|
|
|
|
|
|
|
|
|
|
if (wake_mask & USB_PORT_FEAT_REMOTE_WAKE_CONNECT)
|
|
|
|
|
temp |= PORT_WKCONN_E;
|
|
|
|
|
else
|
|
|
|
|
temp &= ~PORT_WKCONN_E;
|
|
|
|
|
|
|
|
|
|
if (wake_mask & USB_PORT_FEAT_REMOTE_WAKE_DISCONNECT)
|
|
|
|
|
temp |= PORT_WKDISC_E;
|
|
|
|
|
else
|
|
|
|
|
temp &= ~PORT_WKDISC_E;
|
|
|
|
|
|
|
|
|
|
if (wake_mask & USB_PORT_FEAT_REMOTE_WAKE_OVER_CURRENT)
|
|
|
|
|
temp |= PORT_WKOC_E;
|
|
|
|
|
else
|
|
|
|
|
temp &= ~PORT_WKOC_E;
|
|
|
|
|
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[port_id]);
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
}
|
|
|
|
|
|
2011-09-24 04:19:49 +07:00
|
|
|
|
/* Test and clear port RWC bit */
|
|
|
|
|
void xhci_test_and_clear_bit(struct xhci_hcd *xhci, __le32 __iomem **port_array,
|
|
|
|
|
int port_id, u32 port_bit)
|
|
|
|
|
{
|
|
|
|
|
u32 temp;
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[port_id]);
|
2011-09-24 04:19:49 +07:00
|
|
|
|
if (temp & port_bit) {
|
|
|
|
|
temp = xhci_port_state_to_neutral(temp);
|
|
|
|
|
temp |= port_bit;
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[port_id]);
|
2011-09-24 04:19:49 +07:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2013-04-02 23:23:42 +07:00
|
|
|
|
/* Updates Link Status for USB 2.1 port */
|
|
|
|
|
static void xhci_hub_report_usb2_link_state(u32 *status, u32 status_reg)
|
|
|
|
|
{
|
|
|
|
|
if ((status_reg & PORT_PLS_MASK) == XDEV_U2)
|
|
|
|
|
*status |= USB_PORT_STAT_L1;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-18 20:20:00 +07:00
|
|
|
|
/* Updates Link Status for super Speed port */
|
2014-08-28 04:38:04 +07:00
|
|
|
|
static void xhci_hub_report_usb3_link_state(struct xhci_hcd *xhci,
|
|
|
|
|
u32 *status, u32 status_reg)
|
2012-06-18 20:20:00 +07:00
|
|
|
|
{
|
|
|
|
|
u32 pls = status_reg & PORT_PLS_MASK;
|
|
|
|
|
|
|
|
|
|
/* resume state is a xHCI internal state.
|
|
|
|
|
* Do not report it to usb core.
|
|
|
|
|
*/
|
|
|
|
|
if (pls == XDEV_RESUME)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
/* When the CAS bit is set then warm reset
|
|
|
|
|
* should be performed on port
|
|
|
|
|
*/
|
|
|
|
|
if (status_reg & PORT_CAS) {
|
|
|
|
|
/* The CAS bit can be set while the port is
|
|
|
|
|
* in any link state.
|
|
|
|
|
* Only roothubs have CAS bit, so we
|
|
|
|
|
* pretend to be in compliance mode
|
|
|
|
|
* unless we're already in compliance
|
|
|
|
|
* or the inactive state.
|
|
|
|
|
*/
|
|
|
|
|
if (pls != USB_SS_PORT_LS_COMP_MOD &&
|
|
|
|
|
pls != USB_SS_PORT_LS_SS_INACTIVE) {
|
|
|
|
|
pls = USB_SS_PORT_LS_COMP_MOD;
|
|
|
|
|
}
|
|
|
|
|
/* Return also connection bit -
|
|
|
|
|
* hub state machine resets port
|
|
|
|
|
* when this bit is set.
|
|
|
|
|
*/
|
|
|
|
|
pls |= USB_PORT_STAT_CONNECTION;
|
usb: host: xhci: Fix Compliance Mode on SN65LVPE502CP Hardware
This patch is intended to work around a known issue on the
SN65LVPE502CP USB3.0 re-driver that can delay the negotiation
between a device and the host past the usual handshake timeout.
If that happens on the first insertion, the host controller
port will enter in Compliance Mode and NO port status event will
be generated (as per xHCI Spec) making impossible to detect this
event by software. The port will remain in compliance mode until
a warm reset is applied to it.
As a result of this, the port will seem "dead" to the user and no
device connections or disconnections will be detected.
For solving this, the patch creates a timer which polls every 2
seconds the link state of each host controller's port (this
by reading the PORTSC register) and recovers the port by issuing a
Warm reset every time Compliance mode is detected.
If a xHC USB3.0 port has previously entered to U0, the compliance
mode issue will NOT occur only until system resumes from
sleep/hibernate, therefore, the compliance mode timer is stopped
when all xHC USB 3.0 ports have entered U0. The timer is initialized
again after each system resume.
Since the issue is being caused by a piece of hardware, the timer
will be enabled ONLY on those systems that have the SN65LVPE502CP
installed (this patch uses DMI strings for detecting those systems)
therefore making this patch to act as a quirk (XHCI_COMP_MODE_QUIRK
has been added to the xhci stack).
This patch applies for these systems:
Vendor: Hewlett-Packard. System Models: Z420, Z620 and Z820.
This patch should be backported to kernels as old as 3.2, as that was
the first kernel to support warm reset. The kernels will need to
contain both commit 10d674a82e553cb8a1f41027bb3c3e309b3f6804 "USB: When
hot reset for USB3 fails, try warm reset" and commit
8bea2bd37df08aaa599aa361a9f8b836ba98e554 "usb: Add support for root hub
port status CAS". The first patch add warm reset support, and the
second patch modifies the USB core to issue a warm reset when the port
is in compliance mode.
Signed-off-by: Alexis R. Cortes <alexis.cortes@ti.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org
2012-08-04 02:00:27 +07:00
|
|
|
|
} else {
|
|
|
|
|
/*
|
|
|
|
|
* If CAS bit isn't set but the Port is already at
|
|
|
|
|
* Compliance Mode, fake a connection so the USB core
|
|
|
|
|
* notices the Compliance state and resets the port.
|
|
|
|
|
* This resolves an issue generated by the SN65LVPE502CP
|
|
|
|
|
* in which sometimes the port enters compliance mode
|
|
|
|
|
* caused by a delay on the host-device negotiation.
|
|
|
|
|
*/
|
2014-08-28 04:38:04 +07:00
|
|
|
|
if ((xhci->quirks & XHCI_COMP_MODE_QUIRK) &&
|
|
|
|
|
(pls == USB_SS_PORT_LS_COMP_MOD))
|
usb: host: xhci: Fix Compliance Mode on SN65LVPE502CP Hardware
This patch is intended to work around a known issue on the
SN65LVPE502CP USB3.0 re-driver that can delay the negotiation
between a device and the host past the usual handshake timeout.
If that happens on the first insertion, the host controller
port will enter in Compliance Mode and NO port status event will
be generated (as per xHCI Spec) making impossible to detect this
event by software. The port will remain in compliance mode until
a warm reset is applied to it.
As a result of this, the port will seem "dead" to the user and no
device connections or disconnections will be detected.
For solving this, the patch creates a timer which polls every 2
seconds the link state of each host controller's port (this
by reading the PORTSC register) and recovers the port by issuing a
Warm reset every time Compliance mode is detected.
If a xHC USB3.0 port has previously entered to U0, the compliance
mode issue will NOT occur only until system resumes from
sleep/hibernate, therefore, the compliance mode timer is stopped
when all xHC USB 3.0 ports have entered U0. The timer is initialized
again after each system resume.
Since the issue is being caused by a piece of hardware, the timer
will be enabled ONLY on those systems that have the SN65LVPE502CP
installed (this patch uses DMI strings for detecting those systems)
therefore making this patch to act as a quirk (XHCI_COMP_MODE_QUIRK
has been added to the xhci stack).
This patch applies for these systems:
Vendor: Hewlett-Packard. System Models: Z420, Z620 and Z820.
This patch should be backported to kernels as old as 3.2, as that was
the first kernel to support warm reset. The kernels will need to
contain both commit 10d674a82e553cb8a1f41027bb3c3e309b3f6804 "USB: When
hot reset for USB3 fails, try warm reset" and commit
8bea2bd37df08aaa599aa361a9f8b836ba98e554 "usb: Add support for root hub
port status CAS". The first patch add warm reset support, and the
second patch modifies the USB core to issue a warm reset when the port
is in compliance mode.
Signed-off-by: Alexis R. Cortes <alexis.cortes@ti.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org
2012-08-04 02:00:27 +07:00
|
|
|
|
pls |= USB_PORT_STAT_CONNECTION;
|
2012-06-18 20:20:00 +07:00
|
|
|
|
}
|
usb: host: xhci: Fix Compliance Mode on SN65LVPE502CP Hardware
This patch is intended to work around a known issue on the
SN65LVPE502CP USB3.0 re-driver that can delay the negotiation
between a device and the host past the usual handshake timeout.
If that happens on the first insertion, the host controller
port will enter in Compliance Mode and NO port status event will
be generated (as per xHCI Spec) making impossible to detect this
event by software. The port will remain in compliance mode until
a warm reset is applied to it.
As a result of this, the port will seem "dead" to the user and no
device connections or disconnections will be detected.
For solving this, the patch creates a timer which polls every 2
seconds the link state of each host controller's port (this
by reading the PORTSC register) and recovers the port by issuing a
Warm reset every time Compliance mode is detected.
If a xHC USB3.0 port has previously entered to U0, the compliance
mode issue will NOT occur only until system resumes from
sleep/hibernate, therefore, the compliance mode timer is stopped
when all xHC USB 3.0 ports have entered U0. The timer is initialized
again after each system resume.
Since the issue is being caused by a piece of hardware, the timer
will be enabled ONLY on those systems that have the SN65LVPE502CP
installed (this patch uses DMI strings for detecting those systems)
therefore making this patch to act as a quirk (XHCI_COMP_MODE_QUIRK
has been added to the xhci stack).
This patch applies for these systems:
Vendor: Hewlett-Packard. System Models: Z420, Z620 and Z820.
This patch should be backported to kernels as old as 3.2, as that was
the first kernel to support warm reset. The kernels will need to
contain both commit 10d674a82e553cb8a1f41027bb3c3e309b3f6804 "USB: When
hot reset for USB3 fails, try warm reset" and commit
8bea2bd37df08aaa599aa361a9f8b836ba98e554 "usb: Add support for root hub
port status CAS". The first patch add warm reset support, and the
second patch modifies the USB core to issue a warm reset when the port
is in compliance mode.
Signed-off-by: Alexis R. Cortes <alexis.cortes@ti.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org
2012-08-04 02:00:27 +07:00
|
|
|
|
|
2012-06-18 20:20:00 +07:00
|
|
|
|
/* update status field */
|
|
|
|
|
*status |= pls;
|
|
|
|
|
}
|
|
|
|
|
|
usb: host: xhci: Fix Compliance Mode on SN65LVPE502CP Hardware
This patch is intended to work around a known issue on the
SN65LVPE502CP USB3.0 re-driver that can delay the negotiation
between a device and the host past the usual handshake timeout.
If that happens on the first insertion, the host controller
port will enter in Compliance Mode and NO port status event will
be generated (as per xHCI Spec) making impossible to detect this
event by software. The port will remain in compliance mode until
a warm reset is applied to it.
As a result of this, the port will seem "dead" to the user and no
device connections or disconnections will be detected.
For solving this, the patch creates a timer which polls every 2
seconds the link state of each host controller's port (this
by reading the PORTSC register) and recovers the port by issuing a
Warm reset every time Compliance mode is detected.
If a xHC USB3.0 port has previously entered to U0, the compliance
mode issue will NOT occur only until system resumes from
sleep/hibernate, therefore, the compliance mode timer is stopped
when all xHC USB 3.0 ports have entered U0. The timer is initialized
again after each system resume.
Since the issue is being caused by a piece of hardware, the timer
will be enabled ONLY on those systems that have the SN65LVPE502CP
installed (this patch uses DMI strings for detecting those systems)
therefore making this patch to act as a quirk (XHCI_COMP_MODE_QUIRK
has been added to the xhci stack).
This patch applies for these systems:
Vendor: Hewlett-Packard. System Models: Z420, Z620 and Z820.
This patch should be backported to kernels as old as 3.2, as that was
the first kernel to support warm reset. The kernels will need to
contain both commit 10d674a82e553cb8a1f41027bb3c3e309b3f6804 "USB: When
hot reset for USB3 fails, try warm reset" and commit
8bea2bd37df08aaa599aa361a9f8b836ba98e554 "usb: Add support for root hub
port status CAS". The first patch add warm reset support, and the
second patch modifies the USB core to issue a warm reset when the port
is in compliance mode.
Signed-off-by: Alexis R. Cortes <alexis.cortes@ti.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org
2012-08-04 02:00:27 +07:00
|
|
|
|
/*
|
|
|
|
|
* Function for Compliance Mode Quirk.
|
|
|
|
|
*
|
|
|
|
|
* This Function verifies if all xhc USB3 ports have entered U0, if so,
|
|
|
|
|
* the compliance mode timer is deleted. A port won't enter
|
|
|
|
|
* compliance mode if it has previously entered U0.
|
|
|
|
|
*/
|
2013-09-16 13:31:34 +07:00
|
|
|
|
static void xhci_del_comp_mod_timer(struct xhci_hcd *xhci, u32 status,
|
|
|
|
|
u16 wIndex)
|
usb: host: xhci: Fix Compliance Mode on SN65LVPE502CP Hardware
This patch is intended to work around a known issue on the
SN65LVPE502CP USB3.0 re-driver that can delay the negotiation
between a device and the host past the usual handshake timeout.
If that happens on the first insertion, the host controller
port will enter in Compliance Mode and NO port status event will
be generated (as per xHCI Spec) making impossible to detect this
event by software. The port will remain in compliance mode until
a warm reset is applied to it.
As a result of this, the port will seem "dead" to the user and no
device connections or disconnections will be detected.
For solving this, the patch creates a timer which polls every 2
seconds the link state of each host controller's port (this
by reading the PORTSC register) and recovers the port by issuing a
Warm reset every time Compliance mode is detected.
If a xHC USB3.0 port has previously entered to U0, the compliance
mode issue will NOT occur only until system resumes from
sleep/hibernate, therefore, the compliance mode timer is stopped
when all xHC USB 3.0 ports have entered U0. The timer is initialized
again after each system resume.
Since the issue is being caused by a piece of hardware, the timer
will be enabled ONLY on those systems that have the SN65LVPE502CP
installed (this patch uses DMI strings for detecting those systems)
therefore making this patch to act as a quirk (XHCI_COMP_MODE_QUIRK
has been added to the xhci stack).
This patch applies for these systems:
Vendor: Hewlett-Packard. System Models: Z420, Z620 and Z820.
This patch should be backported to kernels as old as 3.2, as that was
the first kernel to support warm reset. The kernels will need to
contain both commit 10d674a82e553cb8a1f41027bb3c3e309b3f6804 "USB: When
hot reset for USB3 fails, try warm reset" and commit
8bea2bd37df08aaa599aa361a9f8b836ba98e554 "usb: Add support for root hub
port status CAS". The first patch add warm reset support, and the
second patch modifies the USB core to issue a warm reset when the port
is in compliance mode.
Signed-off-by: Alexis R. Cortes <alexis.cortes@ti.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org
2012-08-04 02:00:27 +07:00
|
|
|
|
{
|
|
|
|
|
u32 all_ports_seen_u0 = ((1 << xhci->num_usb3_ports)-1);
|
|
|
|
|
bool port_in_u0 = ((status & PORT_PLS_MASK) == XDEV_U0);
|
|
|
|
|
|
|
|
|
|
if (!(xhci->quirks & XHCI_COMP_MODE_QUIRK))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if ((xhci->port_status_u0 != all_ports_seen_u0) && port_in_u0) {
|
|
|
|
|
xhci->port_status_u0 |= 1 << wIndex;
|
|
|
|
|
if (xhci->port_status_u0 == all_ports_seen_u0) {
|
|
|
|
|
del_timer_sync(&xhci->comp_mode_recovery_timer);
|
2013-08-06 11:52:45 +07:00
|
|
|
|
xhci_dbg_trace(xhci, trace_xhci_dbg_quirks,
|
|
|
|
|
"All USB3 ports have entered U0 already!");
|
|
|
|
|
xhci_dbg_trace(xhci, trace_xhci_dbg_quirks,
|
|
|
|
|
"Compliance Mode Recovery Timer Deleted.");
|
usb: host: xhci: Fix Compliance Mode on SN65LVPE502CP Hardware
This patch is intended to work around a known issue on the
SN65LVPE502CP USB3.0 re-driver that can delay the negotiation
between a device and the host past the usual handshake timeout.
If that happens on the first insertion, the host controller
port will enter in Compliance Mode and NO port status event will
be generated (as per xHCI Spec) making impossible to detect this
event by software. The port will remain in compliance mode until
a warm reset is applied to it.
As a result of this, the port will seem "dead" to the user and no
device connections or disconnections will be detected.
For solving this, the patch creates a timer which polls every 2
seconds the link state of each host controller's port (this
by reading the PORTSC register) and recovers the port by issuing a
Warm reset every time Compliance mode is detected.
If a xHC USB3.0 port has previously entered to U0, the compliance
mode issue will NOT occur only until system resumes from
sleep/hibernate, therefore, the compliance mode timer is stopped
when all xHC USB 3.0 ports have entered U0. The timer is initialized
again after each system resume.
Since the issue is being caused by a piece of hardware, the timer
will be enabled ONLY on those systems that have the SN65LVPE502CP
installed (this patch uses DMI strings for detecting those systems)
therefore making this patch to act as a quirk (XHCI_COMP_MODE_QUIRK
has been added to the xhci stack).
This patch applies for these systems:
Vendor: Hewlett-Packard. System Models: Z420, Z620 and Z820.
This patch should be backported to kernels as old as 3.2, as that was
the first kernel to support warm reset. The kernels will need to
contain both commit 10d674a82e553cb8a1f41027bb3c3e309b3f6804 "USB: When
hot reset for USB3 fails, try warm reset" and commit
8bea2bd37df08aaa599aa361a9f8b836ba98e554 "usb: Add support for root hub
port status CAS". The first patch add warm reset support, and the
second patch modifies the USB core to issue a warm reset when the port
is in compliance mode.
Signed-off-by: Alexis R. Cortes <alexis.cortes@ti.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org
2012-08-04 02:00:27 +07:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2013-04-02 22:42:20 +07:00
|
|
|
|
/*
|
|
|
|
|
* Converts a raw xHCI port status into the format that external USB 2.0 or USB
|
|
|
|
|
* 3.0 hubs use.
|
|
|
|
|
*
|
|
|
|
|
* Possible side effects:
|
|
|
|
|
* - Mark a port as being done with device resume,
|
|
|
|
|
* and ring the endpoint doorbells.
|
|
|
|
|
* - Stop the Synopsys redriver Compliance Mode polling.
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
* - Drop and reacquire the xHCI lock, in order to wait for port resume.
|
2013-04-02 22:42:20 +07:00
|
|
|
|
*/
|
|
|
|
|
static u32 xhci_get_port_status(struct usb_hcd *hcd,
|
|
|
|
|
struct xhci_bus_state *bus_state,
|
|
|
|
|
__le32 __iomem **port_array,
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
u16 wIndex, u32 raw_port_status,
|
|
|
|
|
unsigned long flags)
|
|
|
|
|
__releases(&xhci->lock)
|
|
|
|
|
__acquires(&xhci->lock)
|
2013-04-02 22:42:20 +07:00
|
|
|
|
{
|
|
|
|
|
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
|
|
|
|
|
u32 status = 0;
|
|
|
|
|
int slot_id;
|
|
|
|
|
|
|
|
|
|
/* wPortChange bits */
|
|
|
|
|
if (raw_port_status & PORT_CSC)
|
|
|
|
|
status |= USB_PORT_STAT_C_CONNECTION << 16;
|
|
|
|
|
if (raw_port_status & PORT_PEC)
|
|
|
|
|
status |= USB_PORT_STAT_C_ENABLE << 16;
|
|
|
|
|
if ((raw_port_status & PORT_OCC))
|
|
|
|
|
status |= USB_PORT_STAT_C_OVERCURRENT << 16;
|
|
|
|
|
if ((raw_port_status & PORT_RC))
|
|
|
|
|
status |= USB_PORT_STAT_C_RESET << 16;
|
|
|
|
|
/* USB3.0 only */
|
|
|
|
|
if (hcd->speed == HCD_USB3) {
|
|
|
|
|
if ((raw_port_status & PORT_PLC))
|
|
|
|
|
status |= USB_PORT_STAT_C_LINK_STATE << 16;
|
|
|
|
|
if ((raw_port_status & PORT_WRC))
|
|
|
|
|
status |= USB_PORT_STAT_C_BH_RESET << 16;
|
2015-03-23 23:27:41 +07:00
|
|
|
|
if ((raw_port_status & PORT_CEC))
|
|
|
|
|
status |= USB_PORT_STAT_C_CONFIG_ERROR << 16;
|
2013-04-02 22:42:20 +07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (hcd->speed != HCD_USB3) {
|
|
|
|
|
if ((raw_port_status & PORT_PLS_MASK) == XDEV_U3
|
|
|
|
|
&& (raw_port_status & PORT_POWER))
|
|
|
|
|
status |= USB_PORT_STAT_SUSPEND;
|
|
|
|
|
}
|
|
|
|
|
if ((raw_port_status & PORT_PLS_MASK) == XDEV_RESUME &&
|
|
|
|
|
!DEV_SUPERSPEED(raw_port_status)) {
|
|
|
|
|
if ((raw_port_status & PORT_RESET) ||
|
|
|
|
|
!(raw_port_status & PORT_PE))
|
|
|
|
|
return 0xffffffff;
|
|
|
|
|
if (time_after_eq(jiffies,
|
|
|
|
|
bus_state->resume_done[wIndex])) {
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
int time_left;
|
|
|
|
|
|
2013-04-02 22:42:20 +07:00
|
|
|
|
xhci_dbg(xhci, "Resume USB2 port %d\n",
|
|
|
|
|
wIndex + 1);
|
|
|
|
|
bus_state->resume_done[wIndex] = 0;
|
|
|
|
|
clear_bit(wIndex, &bus_state->resuming_ports);
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
|
|
|
|
|
set_bit(wIndex, &bus_state->rexit_ports);
|
2013-04-02 22:42:20 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex,
|
|
|
|
|
XDEV_U0);
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
time_left = wait_for_completion_timeout(
|
|
|
|
|
&bus_state->rexit_done[wIndex],
|
|
|
|
|
msecs_to_jiffies(
|
|
|
|
|
XHCI_MAX_REXIT_TIMEOUT));
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
|
|
|
|
if (time_left) {
|
|
|
|
|
slot_id = xhci_find_slot_id_by_port(hcd,
|
|
|
|
|
xhci, wIndex + 1);
|
|
|
|
|
if (!slot_id) {
|
|
|
|
|
xhci_dbg(xhci, "slot_id is zero\n");
|
|
|
|
|
return 0xffffffff;
|
|
|
|
|
}
|
|
|
|
|
xhci_ring_device(xhci, slot_id);
|
|
|
|
|
} else {
|
2013-11-15 10:34:06 +07:00
|
|
|
|
int port_status = readl(port_array[wIndex]);
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
xhci_warn(xhci, "Port resume took longer than %i msec, port status = 0x%x\n",
|
|
|
|
|
XHCI_MAX_REXIT_TIMEOUT,
|
|
|
|
|
port_status);
|
|
|
|
|
status |= USB_PORT_STAT_SUSPEND;
|
|
|
|
|
clear_bit(wIndex, &bus_state->rexit_ports);
|
2013-04-02 22:42:20 +07:00
|
|
|
|
}
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
|
2013-04-02 22:42:20 +07:00
|
|
|
|
bus_state->port_c_suspend |= 1 << wIndex;
|
|
|
|
|
bus_state->suspended_ports &= ~(1 << wIndex);
|
|
|
|
|
} else {
|
|
|
|
|
/*
|
|
|
|
|
* The resume has been signaling for less than
|
|
|
|
|
* 20ms. Report the port status as SUSPEND,
|
|
|
|
|
* let the usbcore check port status again
|
|
|
|
|
* and clear resume signaling later.
|
|
|
|
|
*/
|
|
|
|
|
status |= USB_PORT_STAT_SUSPEND;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
if ((raw_port_status & PORT_PLS_MASK) == XDEV_U0
|
|
|
|
|
&& (raw_port_status & PORT_POWER)
|
|
|
|
|
&& (bus_state->suspended_ports & (1 << wIndex))) {
|
|
|
|
|
bus_state->suspended_ports &= ~(1 << wIndex);
|
|
|
|
|
if (hcd->speed != HCD_USB3)
|
|
|
|
|
bus_state->port_c_suspend |= 1 << wIndex;
|
|
|
|
|
}
|
|
|
|
|
if (raw_port_status & PORT_CONNECT) {
|
|
|
|
|
status |= USB_PORT_STAT_CONNECTION;
|
|
|
|
|
status |= xhci_port_speed(raw_port_status);
|
|
|
|
|
}
|
|
|
|
|
if (raw_port_status & PORT_PE)
|
|
|
|
|
status |= USB_PORT_STAT_ENABLE;
|
|
|
|
|
if (raw_port_status & PORT_OC)
|
|
|
|
|
status |= USB_PORT_STAT_OVERCURRENT;
|
|
|
|
|
if (raw_port_status & PORT_RESET)
|
|
|
|
|
status |= USB_PORT_STAT_RESET;
|
|
|
|
|
if (raw_port_status & PORT_POWER) {
|
|
|
|
|
if (hcd->speed == HCD_USB3)
|
|
|
|
|
status |= USB_SS_PORT_STAT_POWER;
|
|
|
|
|
else
|
|
|
|
|
status |= USB_PORT_STAT_POWER;
|
|
|
|
|
}
|
2013-04-02 23:23:42 +07:00
|
|
|
|
/* Update Port Link State */
|
2013-04-02 22:42:20 +07:00
|
|
|
|
if (hcd->speed == HCD_USB3) {
|
2014-08-28 04:38:04 +07:00
|
|
|
|
xhci_hub_report_usb3_link_state(xhci, &status, raw_port_status);
|
2013-04-02 22:42:20 +07:00
|
|
|
|
/*
|
|
|
|
|
* Verify if all USB3 Ports Have entered U0 already.
|
|
|
|
|
* Delete Compliance Mode Timer if so.
|
|
|
|
|
*/
|
|
|
|
|
xhci_del_comp_mod_timer(xhci, raw_port_status, wIndex);
|
2013-04-02 23:23:42 +07:00
|
|
|
|
} else {
|
|
|
|
|
xhci_hub_report_usb2_link_state(&status, raw_port_status);
|
2013-04-02 22:42:20 +07:00
|
|
|
|
}
|
|
|
|
|
if (bus_state->port_c_suspend & (1 << wIndex))
|
|
|
|
|
status |= 1 << USB_PORT_FEAT_C_SUSPEND;
|
|
|
|
|
|
|
|
|
|
return status;
|
|
|
|
|
}
|
|
|
|
|
|
2009-04-28 09:57:12 +07:00
|
|
|
|
int xhci_hub_control(struct usb_hcd *hcd, u16 typeReq, u16 wValue,
|
|
|
|
|
u16 wIndex, char *buf, u16 wLength)
|
|
|
|
|
{
|
|
|
|
|
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
|
2011-05-03 20:11:00 +07:00
|
|
|
|
int max_ports;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
unsigned long flags;
|
2011-09-24 04:19:48 +07:00
|
|
|
|
u32 temp, status;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
int retval = 0;
|
2011-03-29 09:40:46 +07:00
|
|
|
|
__le32 __iomem **port_array;
|
2010-10-14 21:22:57 +07:00
|
|
|
|
int slot_id;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
struct xhci_bus_state *bus_state;
|
2011-04-27 17:07:39 +07:00
|
|
|
|
u16 link_state = 0;
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
u16 wake_mask = 0;
|
2011-11-11 07:02:13 +07:00
|
|
|
|
u16 timeout = 0;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
2011-05-03 20:11:00 +07:00
|
|
|
|
max_ports = xhci_get_ports(hcd, &port_array);
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state = &xhci->bus_state[hcd_index(hcd)];
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
switch (typeReq) {
|
|
|
|
|
case GetHubStatus:
|
|
|
|
|
/* No power source, over-current reported per port */
|
|
|
|
|
memset(buf, 0, 4);
|
|
|
|
|
break;
|
|
|
|
|
case GetHubDescriptor:
|
2010-11-30 07:14:37 +07:00
|
|
|
|
/* Check to make sure userspace is asking for the USB 3.0 hub
|
|
|
|
|
* descriptor for the USB 3.0 roothub. If not, we stall the
|
|
|
|
|
* endpoint, like external hubs do.
|
|
|
|
|
*/
|
|
|
|
|
if (hcd->speed == HCD_USB3 &&
|
|
|
|
|
(wLength < USB_DT_SS_HUB_SIZE ||
|
|
|
|
|
wValue != (USB_DT_SS_HUB << 8))) {
|
|
|
|
|
xhci_dbg(xhci, "Wrong hub descriptor type for "
|
|
|
|
|
"USB 3.0 roothub.\n");
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
xhci_hub_descriptor(hcd, xhci,
|
|
|
|
|
(struct usb_hub_descriptor *) buf);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
break;
|
2011-10-07 01:54:23 +07:00
|
|
|
|
case DeviceRequest | USB_REQ_GET_DESCRIPTOR:
|
|
|
|
|
if ((wValue & 0xff00) != (USB_DT_BOS << 8))
|
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
|
|
if (hcd->speed != HCD_USB3)
|
|
|
|
|
goto error;
|
|
|
|
|
|
2012-06-25 22:24:30 +07:00
|
|
|
|
/* Set the U1 and U2 exit latencies. */
|
2011-10-07 01:54:23 +07:00
|
|
|
|
memcpy(buf, &usb_bos_descriptor,
|
|
|
|
|
USB_DT_BOS_SIZE + USB_DT_USB_SS_CAP_SIZE);
|
usb/xhci: Change how we indicate a host supports Link PM.
The xHCI driver currently uses a USB core internal field,
udev->lpm_capable, to indicate the xHCI driver knows how to calculate
the LPM timeout values. If this value is set for the host controller
udev, it means Link PM can be enabled for child devices under that host.
Change the code so the xHCI driver isn't mucking with USB core internal
fields. Instead, indicate the xHCI driver doesn't support Link PM on
this host by clearing the U1 and U2 exit latencies in the roothub
SuperSpeed Extended Capabilities BOS descriptor.
The code to check for the roothub setting U1 and U2 exit latencies to
zero will also disable LPM for external devices that do that same. This
was already effectively done with commit
ae8963adb4ad8c5f2a89ca1d99fb7bb721e7599f "usb: Don't enable LPM if the
exit latency is zero." Leave that code in place, so that if a device
sets one exit latency value to zero, but the other is set to a valid
value, LPM is only enabled for the U1 or U2 state that had the valid
value. This is the same behavior the code had before.
Also, change messages about missing Link PM information from warning
level to info level. Only print a warning about the first device that
doesn't support LPM, to avoid log spam. Further, cleanup some
unnecessary line breaks to help people to grep for the error messages.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
2014-01-18 05:15:44 +07:00
|
|
|
|
if ((xhci->quirks & XHCI_LPM_SUPPORT)) {
|
|
|
|
|
temp = readl(&xhci->cap_regs->hcs_params3);
|
|
|
|
|
buf[12] = HCS_U1_LATENCY(temp);
|
|
|
|
|
put_unaligned_le16(HCS_U2_LATENCY(temp), &buf[13]);
|
|
|
|
|
}
|
2011-10-07 01:54:23 +07:00
|
|
|
|
|
2012-06-25 22:24:30 +07:00
|
|
|
|
/* Indicate whether the host has LTM support. */
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(&xhci->cap_regs->hcc_params);
|
2012-06-25 22:24:30 +07:00
|
|
|
|
if (HCC_LTC(temp))
|
|
|
|
|
buf[8] |= USB_LTM_SUPPORT;
|
|
|
|
|
|
2011-10-07 01:54:23 +07:00
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
return USB_DT_BOS_SIZE + USB_DT_USB_SS_CAP_SIZE;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
case GetPortStatus:
|
2011-05-03 20:11:00 +07:00
|
|
|
|
if (!wIndex || wIndex > max_ports)
|
2009-04-28 09:57:12 +07:00
|
|
|
|
goto error;
|
|
|
|
|
wIndex--;
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2010-10-30 04:37:23 +07:00
|
|
|
|
if (temp == 0xffffffff) {
|
|
|
|
|
retval = -ENODEV;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2013-04-02 22:42:20 +07:00
|
|
|
|
status = xhci_get_port_status(hcd, bus_state, port_array,
|
usb: Fix xHCI host issues on remote wakeup.
When a device signals remote wakeup on a roothub, and the suspend change
bit is set, the host controller driver must not give control back to the
USB core until the port goes back into the active state.
EHCI accomplishes this by waiting in the get port status function until
the PORT_RESUME bit is cleared:
/* stop resume signaling */
temp &= ~(PORT_RWC_BITS | PORT_SUSPEND | PORT_RESUME);
ehci_writel(ehci, temp, status_reg);
clear_bit(wIndex, &ehci->resuming_ports);
retval = ehci_handshake(ehci, status_reg,
PORT_RESUME, 0, 2000 /* 2msec */);
Similarly, the xHCI host should wait until the port goes into U0, before
passing control up to the USB core. When the port transitions from the
RExit state to U0, the xHCI driver will get a port status change event.
We need to wait for that event before passing control up to the USB
core.
After the port transitions to the active state, the USB core should time
a recovery interval before it talks to the device. The length of that
recovery interval is TRSMRCY, 10 ms, mentioned in the USB 2.0 spec,
section 7.1.7.7. The previous xHCI code (which did not wait for the
port to go into U0) would cause the USB core to violate that recovery
interval.
This bug caused numerous USB device disconnects on remote wakeup under
ChromeOS and a Lynx Point LP xHCI host that takes up to 20 ms to move
from RExit to U0. ChromeOS is very aggressive about power savings, and
sets the autosuspend_delay to 100 ms, and disables USB persist.
I attempted to replicate this bug with Ubuntu 12.04, but could not. I
used Ubuntu 12.04 on the same platform, with the same BIOS that the bug
was triggered on ChromeOS with. I also changed the USB sysfs settings
as described above, but still could not reproduce the bug under Ubuntu.
It may be that ChromeOS userspace triggers this bug through additional
settings.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2013-08-20 22:12:12 +07:00
|
|
|
|
wIndex, temp, flags);
|
2013-04-02 22:42:20 +07:00
|
|
|
|
if (status == 0xffffffff)
|
|
|
|
|
goto error;
|
2011-04-27 17:07:43 +07:00
|
|
|
|
|
2013-04-02 22:42:20 +07:00
|
|
|
|
xhci_dbg(xhci, "get port status, actual port %d status = 0x%x\n",
|
|
|
|
|
wIndex, temp);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
xhci_dbg(xhci, "Get port status returned 0x%x\n", status);
|
2013-04-02 22:42:20 +07:00
|
|
|
|
|
2009-04-28 09:57:12 +07:00
|
|
|
|
put_unaligned(cpu_to_le32(status), (__le32 *) buf);
|
|
|
|
|
break;
|
|
|
|
|
case SetPortFeature:
|
2011-04-27 17:07:39 +07:00
|
|
|
|
if (wValue == USB_PORT_FEAT_LINK_STATE)
|
|
|
|
|
link_state = (wIndex & 0xff00) >> 3;
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
if (wValue == USB_PORT_FEAT_REMOTE_WAKE_MASK)
|
|
|
|
|
wake_mask = wIndex & 0xff00;
|
2011-11-11 07:02:13 +07:00
|
|
|
|
/* The MSB of wIndex is the U1/U2 timeout */
|
|
|
|
|
timeout = (wIndex & 0xff00) >> 8;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
wIndex &= 0xff;
|
2011-05-03 20:11:00 +07:00
|
|
|
|
if (!wIndex || wIndex > max_ports)
|
2009-04-28 09:57:12 +07:00
|
|
|
|
goto error;
|
|
|
|
|
wIndex--;
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2010-10-30 04:37:23 +07:00
|
|
|
|
if (temp == 0xffffffff) {
|
|
|
|
|
retval = -ENODEV;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2009-04-28 09:57:12 +07:00
|
|
|
|
temp = xhci_port_state_to_neutral(temp);
|
2010-11-30 07:14:37 +07:00
|
|
|
|
/* FIXME: What new port features do we need to support? */
|
2009-04-28 09:57:12 +07:00
|
|
|
|
switch (wValue) {
|
2010-10-14 21:22:57 +07:00
|
|
|
|
case USB_PORT_FEAT_SUSPEND:
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2011-09-24 04:19:52 +07:00
|
|
|
|
if ((temp & PORT_PLS_MASK) != XDEV_U0) {
|
|
|
|
|
/* Resume the port to U0 first */
|
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex,
|
|
|
|
|
XDEV_U0);
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
msleep(10);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
}
|
2010-10-14 21:22:57 +07:00
|
|
|
|
/* In spec software should not attempt to suspend
|
|
|
|
|
* a port unless the port reports that it is in the
|
|
|
|
|
* enabled (PED = ‘1’,PLS < ‘3’) state.
|
|
|
|
|
*/
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
if ((temp & PORT_PE) == 0 || (temp & PORT_RESET)
|
|
|
|
|
|| (temp & PORT_PLS_MASK) >= XDEV_U3) {
|
|
|
|
|
xhci_warn(xhci, "USB core suspending device "
|
|
|
|
|
"not in U0/U1/U2.\n");
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
|
2010-12-17 01:49:09 +07:00
|
|
|
|
slot_id = xhci_find_slot_id_by_port(hcd, xhci,
|
|
|
|
|
wIndex + 1);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
if (!slot_id) {
|
|
|
|
|
xhci_warn(xhci, "slot_id is zero\n");
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
/* unlock to execute stop endpoint commands */
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
xhci_stop_device(xhci, slot_id, 1);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex, XDEV_U3);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
msleep(10); /* wait device to enter */
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state->suspended_ports |= 1 << wIndex;
|
2010-10-14 21:22:57 +07:00
|
|
|
|
break;
|
2011-04-27 17:07:39 +07:00
|
|
|
|
case USB_PORT_FEAT_LINK_STATE:
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
USB: Allow USB 3.0 ports to be disabled.
If hot and warm reset fails, or a port remains in the Compliance Mode,
the USB core needs to be able to disable a USB 3.0 port. Unlike USB 2.0
ports, once the port is placed into the Disabled link state, it will not
report any new device connects. To get device connect notifications, we
need to put the link into the Disabled state, and then the RxDetect
state.
The xHCI driver needs to atomically clear all change bits on USB 3.0
port disable, so that we get Port Status Change Events for future port
changes. We could technically do this in the USB core instead of in the
xHCI roothub code, since the port state machine can't advance out of the
disabled state until we set the link state to RxDetect. However,
external USB 3.0 hubs don't need this code. They are level-triggered,
not edge-triggered like xHCI, so they will continue to send interrupt
events when any change bit is set. Therefore it doesn't make sense to
put this code in the USB core.
This patch is part of a series to fix several reports of infinite loops
on device enumeration failure. This includes John, when he boots with
a USB 3.0 device (Roseweil eusb3 enclosure) attached to his NEC 0.96
host controller. The fix requires warm reset support, so it does not
make sense to backport this patch to stable kernels without warm reset
support.
This patch should be backported to kernels as old as 3.2, contain the
commit ID 75d7cf72ab9fa01dc70877aa5c68e8ef477229dc "usbcore: refine warm
reset logic"
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: John Covici <covici@ccs.covici.com>
Cc: stable@vger.kernel.org
2012-11-15 07:42:32 +07:00
|
|
|
|
|
|
|
|
|
/* Disable port */
|
|
|
|
|
if (link_state == USB_SS_PORT_LS_SS_DISABLED) {
|
|
|
|
|
xhci_dbg(xhci, "Disable port %d\n", wIndex);
|
|
|
|
|
temp = xhci_port_state_to_neutral(temp);
|
|
|
|
|
/*
|
|
|
|
|
* Clear all change bits, so that we get a new
|
|
|
|
|
* connection event.
|
|
|
|
|
*/
|
|
|
|
|
temp |= PORT_CSC | PORT_PEC | PORT_WRC |
|
|
|
|
|
PORT_OCC | PORT_RC | PORT_PLC |
|
|
|
|
|
PORT_CEC;
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp | PORT_PE, port_array[wIndex]);
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
USB: Allow USB 3.0 ports to be disabled.
If hot and warm reset fails, or a port remains in the Compliance Mode,
the USB core needs to be able to disable a USB 3.0 port. Unlike USB 2.0
ports, once the port is placed into the Disabled link state, it will not
report any new device connects. To get device connect notifications, we
need to put the link into the Disabled state, and then the RxDetect
state.
The xHCI driver needs to atomically clear all change bits on USB 3.0
port disable, so that we get Port Status Change Events for future port
changes. We could technically do this in the USB core instead of in the
xHCI roothub code, since the port state machine can't advance out of the
disabled state until we set the link state to RxDetect. However,
external USB 3.0 hubs don't need this code. They are level-triggered,
not edge-triggered like xHCI, so they will continue to send interrupt
events when any change bit is set. Therefore it doesn't make sense to
put this code in the USB core.
This patch is part of a series to fix several reports of infinite loops
on device enumeration failure. This includes John, when he boots with
a USB 3.0 device (Roseweil eusb3 enclosure) attached to his NEC 0.96
host controller. The fix requires warm reset support, so it does not
make sense to backport this patch to stable kernels without warm reset
support.
This patch should be backported to kernels as old as 3.2, contain the
commit ID 75d7cf72ab9fa01dc70877aa5c68e8ef477229dc "usbcore: refine warm
reset logic"
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: John Covici <covici@ccs.covici.com>
Cc: stable@vger.kernel.org
2012-11-15 07:42:32 +07:00
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Put link in RxDetect (enable port) */
|
|
|
|
|
if (link_state == USB_SS_PORT_LS_RX_DETECT) {
|
|
|
|
|
xhci_dbg(xhci, "Enable port %d\n", wIndex);
|
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex,
|
|
|
|
|
link_state);
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
USB: Allow USB 3.0 ports to be disabled.
If hot and warm reset fails, or a port remains in the Compliance Mode,
the USB core needs to be able to disable a USB 3.0 port. Unlike USB 2.0
ports, once the port is placed into the Disabled link state, it will not
report any new device connects. To get device connect notifications, we
need to put the link into the Disabled state, and then the RxDetect
state.
The xHCI driver needs to atomically clear all change bits on USB 3.0
port disable, so that we get Port Status Change Events for future port
changes. We could technically do this in the USB core instead of in the
xHCI roothub code, since the port state machine can't advance out of the
disabled state until we set the link state to RxDetect. However,
external USB 3.0 hubs don't need this code. They are level-triggered,
not edge-triggered like xHCI, so they will continue to send interrupt
events when any change bit is set. Therefore it doesn't make sense to
put this code in the USB core.
This patch is part of a series to fix several reports of infinite loops
on device enumeration failure. This includes John, when he boots with
a USB 3.0 device (Roseweil eusb3 enclosure) attached to his NEC 0.96
host controller. The fix requires warm reset support, so it does not
make sense to backport this patch to stable kernels without warm reset
support.
This patch should be backported to kernels as old as 3.2, contain the
commit ID 75d7cf72ab9fa01dc70877aa5c68e8ef477229dc "usbcore: refine warm
reset logic"
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: John Covici <covici@ccs.covici.com>
Cc: stable@vger.kernel.org
2012-11-15 07:42:32 +07:00
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-27 17:07:39 +07:00
|
|
|
|
/* Software should not attempt to set
|
USB: Allow USB 3.0 ports to be disabled.
If hot and warm reset fails, or a port remains in the Compliance Mode,
the USB core needs to be able to disable a USB 3.0 port. Unlike USB 2.0
ports, once the port is placed into the Disabled link state, it will not
report any new device connects. To get device connect notifications, we
need to put the link into the Disabled state, and then the RxDetect
state.
The xHCI driver needs to atomically clear all change bits on USB 3.0
port disable, so that we get Port Status Change Events for future port
changes. We could technically do this in the USB core instead of in the
xHCI roothub code, since the port state machine can't advance out of the
disabled state until we set the link state to RxDetect. However,
external USB 3.0 hubs don't need this code. They are level-triggered,
not edge-triggered like xHCI, so they will continue to send interrupt
events when any change bit is set. Therefore it doesn't make sense to
put this code in the USB core.
This patch is part of a series to fix several reports of infinite loops
on device enumeration failure. This includes John, when he boots with
a USB 3.0 device (Roseweil eusb3 enclosure) attached to his NEC 0.96
host controller. The fix requires warm reset support, so it does not
make sense to backport this patch to stable kernels without warm reset
support.
This patch should be backported to kernels as old as 3.2, contain the
commit ID 75d7cf72ab9fa01dc70877aa5c68e8ef477229dc "usbcore: refine warm
reset logic"
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: John Covici <covici@ccs.covici.com>
Cc: stable@vger.kernel.org
2012-11-15 07:42:32 +07:00
|
|
|
|
* port link state above '3' (U3) and the port
|
2011-04-27 17:07:39 +07:00
|
|
|
|
* must be enabled.
|
|
|
|
|
*/
|
|
|
|
|
if ((temp & PORT_PE) == 0 ||
|
USB: Allow USB 3.0 ports to be disabled.
If hot and warm reset fails, or a port remains in the Compliance Mode,
the USB core needs to be able to disable a USB 3.0 port. Unlike USB 2.0
ports, once the port is placed into the Disabled link state, it will not
report any new device connects. To get device connect notifications, we
need to put the link into the Disabled state, and then the RxDetect
state.
The xHCI driver needs to atomically clear all change bits on USB 3.0
port disable, so that we get Port Status Change Events for future port
changes. We could technically do this in the USB core instead of in the
xHCI roothub code, since the port state machine can't advance out of the
disabled state until we set the link state to RxDetect. However,
external USB 3.0 hubs don't need this code. They are level-triggered,
not edge-triggered like xHCI, so they will continue to send interrupt
events when any change bit is set. Therefore it doesn't make sense to
put this code in the USB core.
This patch is part of a series to fix several reports of infinite loops
on device enumeration failure. This includes John, when he boots with
a USB 3.0 device (Roseweil eusb3 enclosure) attached to his NEC 0.96
host controller. The fix requires warm reset support, so it does not
make sense to backport this patch to stable kernels without warm reset
support.
This patch should be backported to kernels as old as 3.2, contain the
commit ID 75d7cf72ab9fa01dc70877aa5c68e8ef477229dc "usbcore: refine warm
reset logic"
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: John Covici <covici@ccs.covici.com>
Cc: stable@vger.kernel.org
2012-11-15 07:42:32 +07:00
|
|
|
|
(link_state > USB_SS_PORT_LS_U3)) {
|
2011-04-27 17:07:39 +07:00
|
|
|
|
xhci_warn(xhci, "Cannot set link state.\n");
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (link_state == USB_SS_PORT_LS_U3) {
|
|
|
|
|
slot_id = xhci_find_slot_id_by_port(hcd, xhci,
|
|
|
|
|
wIndex + 1);
|
|
|
|
|
if (slot_id) {
|
|
|
|
|
/* unlock to execute stop endpoint
|
|
|
|
|
* commands */
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock,
|
|
|
|
|
flags);
|
|
|
|
|
xhci_stop_device(xhci, slot_id, 1);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex,
|
|
|
|
|
link_state);
|
2011-04-27 17:07:39 +07:00
|
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
msleep(20); /* wait device to enter */
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2011-04-27 17:07:39 +07:00
|
|
|
|
if (link_state == USB_SS_PORT_LS_U3)
|
|
|
|
|
bus_state->suspended_ports |= 1 << wIndex;
|
|
|
|
|
break;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
case USB_PORT_FEAT_POWER:
|
|
|
|
|
/*
|
|
|
|
|
* Turn on ports, even if there isn't per-port switching.
|
|
|
|
|
* HC will report connect events even before this is set.
|
2014-09-19 22:32:23 +07:00
|
|
|
|
* However, hub_wq will ignore the roothub events until
|
2009-04-28 09:57:12 +07:00
|
|
|
|
* the roothub is registered.
|
|
|
|
|
*/
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp | PORT_POWER, port_array[wIndex]);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
xhci_dbg(xhci, "set port power, actual port %d status = 0x%x\n", wIndex, temp);
|
2012-09-05 12:44:36 +07:00
|
|
|
|
|
2012-10-15 14:38:34 +07:00
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
2012-09-05 12:44:36 +07:00
|
|
|
|
temp = usb_acpi_power_manageable(hcd->self.root_hub,
|
|
|
|
|
wIndex);
|
|
|
|
|
if (temp)
|
|
|
|
|
usb_acpi_set_power_state(hcd->self.root_hub,
|
|
|
|
|
wIndex, true);
|
2012-10-15 14:38:34 +07:00
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
break;
|
|
|
|
|
case USB_PORT_FEAT_RESET:
|
|
|
|
|
temp = (temp | PORT_RESET);
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[wIndex]);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
xhci_dbg(xhci, "set port reset, actual port %d status = 0x%x\n", wIndex, temp);
|
|
|
|
|
break;
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
case USB_PORT_FEAT_REMOTE_WAKE_MASK:
|
|
|
|
|
xhci_set_remote_wake_mask(xhci, port_array,
|
|
|
|
|
wIndex, wake_mask);
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
xhci_dbg(xhci, "set port remote wake mask, "
|
|
|
|
|
"actual port %d status = 0x%x\n",
|
|
|
|
|
wIndex, temp);
|
|
|
|
|
break;
|
2011-04-27 17:07:29 +07:00
|
|
|
|
case USB_PORT_FEAT_BH_PORT_RESET:
|
|
|
|
|
temp |= PORT_WR;
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[wIndex]);
|
2011-04-27 17:07:29 +07:00
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2011-04-27 17:07:29 +07:00
|
|
|
|
break;
|
2011-11-11 07:02:13 +07:00
|
|
|
|
case USB_PORT_FEAT_U1_TIMEOUT:
|
|
|
|
|
if (hcd->speed != HCD_USB3)
|
|
|
|
|
goto error;
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex] + PORTPMSC);
|
2011-11-11 07:02:13 +07:00
|
|
|
|
temp &= ~PORT_U1_TIMEOUT_MASK;
|
|
|
|
|
temp |= PORT_U1_TIMEOUT(timeout);
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[wIndex] + PORTPMSC);
|
2011-11-11 07:02:13 +07:00
|
|
|
|
break;
|
|
|
|
|
case USB_PORT_FEAT_U2_TIMEOUT:
|
|
|
|
|
if (hcd->speed != HCD_USB3)
|
|
|
|
|
goto error;
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex] + PORTPMSC);
|
2011-11-11 07:02:13 +07:00
|
|
|
|
temp &= ~PORT_U2_TIMEOUT_MASK;
|
|
|
|
|
temp |= PORT_U2_TIMEOUT(timeout);
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[wIndex] + PORTPMSC);
|
2011-11-11 07:02:13 +07:00
|
|
|
|
break;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
default:
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
2010-12-02 02:34:59 +07:00
|
|
|
|
/* unblock any posted writes */
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
break;
|
|
|
|
|
case ClearPortFeature:
|
2011-05-03 20:11:00 +07:00
|
|
|
|
if (!wIndex || wIndex > max_ports)
|
2009-04-28 09:57:12 +07:00
|
|
|
|
goto error;
|
|
|
|
|
wIndex--;
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2010-10-30 04:37:23 +07:00
|
|
|
|
if (temp == 0xffffffff) {
|
|
|
|
|
retval = -ENODEV;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2010-11-30 07:14:37 +07:00
|
|
|
|
/* FIXME: What new port features do we need to support? */
|
2009-04-28 09:57:12 +07:00
|
|
|
|
temp = xhci_port_state_to_neutral(temp);
|
|
|
|
|
switch (wValue) {
|
2010-10-14 21:22:57 +07:00
|
|
|
|
case USB_PORT_FEAT_SUSPEND:
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[wIndex]);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
xhci_dbg(xhci, "clear USB_PORT_FEAT_SUSPEND\n");
|
|
|
|
|
xhci_dbg(xhci, "PORTSC %04x\n", temp);
|
|
|
|
|
if (temp & PORT_RESET)
|
|
|
|
|
goto error;
|
2011-08-03 15:46:48 +07:00
|
|
|
|
if ((temp & PORT_PLS_MASK) == XDEV_U3) {
|
2010-10-14 21:22:57 +07:00
|
|
|
|
if ((temp & PORT_PE) == 0)
|
|
|
|
|
goto error;
|
|
|
|
|
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex,
|
|
|
|
|
XDEV_RESUME);
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
2011-04-27 17:07:50 +07:00
|
|
|
|
msleep(20);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array, wIndex,
|
|
|
|
|
XDEV_U0);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
}
|
2011-04-27 17:07:50 +07:00
|
|
|
|
bus_state->port_c_suspend |= 1 << wIndex;
|
2010-10-14 21:22:57 +07:00
|
|
|
|
|
2010-12-17 01:49:09 +07:00
|
|
|
|
slot_id = xhci_find_slot_id_by_port(hcd, xhci,
|
|
|
|
|
wIndex + 1);
|
2010-10-14 21:22:57 +07:00
|
|
|
|
if (!slot_id) {
|
|
|
|
|
xhci_dbg(xhci, "slot_id is zero\n");
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
xhci_ring_device(xhci, slot_id);
|
|
|
|
|
break;
|
|
|
|
|
case USB_PORT_FEAT_C_SUSPEND:
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state->port_c_suspend &= ~(1 << wIndex);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
case USB_PORT_FEAT_C_RESET:
|
2011-04-27 17:07:29 +07:00
|
|
|
|
case USB_PORT_FEAT_C_BH_PORT_RESET:
|
2009-04-28 09:57:12 +07:00
|
|
|
|
case USB_PORT_FEAT_C_CONNECTION:
|
|
|
|
|
case USB_PORT_FEAT_C_OVER_CURRENT:
|
2009-12-10 06:59:11 +07:00
|
|
|
|
case USB_PORT_FEAT_C_ENABLE:
|
2011-04-27 17:07:35 +07:00
|
|
|
|
case USB_PORT_FEAT_C_PORT_LINK_STATE:
|
2015-03-23 23:27:41 +07:00
|
|
|
|
case USB_PORT_FEAT_C_PORT_CONFIG_ERROR:
|
2009-12-10 06:59:08 +07:00
|
|
|
|
xhci_clear_port_change_bit(xhci, wValue, wIndex,
|
2010-12-02 02:34:59 +07:00
|
|
|
|
port_array[wIndex], temp);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
break;
|
2009-12-10 06:59:11 +07:00
|
|
|
|
case USB_PORT_FEAT_ENABLE:
|
xhci: Register second xHCI roothub.
This patch changes the xHCI driver to allocate two roothubs. This touches
the driver initialization and shutdown paths, roothub emulation code, and
port status change event handlers. This is a rather large patch, but it
can't be broken up, or it would break git-bisect.
Make the xHCI driver register its own PCI probe function. This will call
the USB core to create the USB 2.0 roothub, and then create the USB 3.0
roothub. This gets the code for registering a shared roothub out of the
USB core, and allows other HCDs later to decide if and how many shared
roothubs they want to allocate.
Make sure the xHCI's reset method marks the xHCI host controller's primary
roothub as the USB 2.0 roothub. This ensures that the high speed bus will
be processed first when the PCI device is resumed, and any USB 3.0 devices
that have migrated over to high speed will migrate back after being reset.
This ensures that USB persist works with these odd devices.
The reset method will also mark the xHCI USB2 roothub as having an
integrated TT. Like EHCI host controllers with a "rate matching hub" the
xHCI USB 2.0 roothub doesn't have an OHCI or UHCI companion controller.
It doesn't really have a TT, but we'll lie and say it has an integrated
TT. We need to do this because the USB core will reject LS/FS devices
under a HS hub without a TT.
Other details:
-------------
The roothub emulation code is changed to return the correct number of
ports for the two roothubs. For the USB 3.0 roothub, it only reports the
USB 3.0 ports. For the USB 2.0 roothub, it reports all the LS/FS/HS
ports. The code to disable a port now checks the speed of the roothub,
and refuses to disable SuperSpeed ports under the USB 3.0 roothub.
The code for initializing a new device context must be changed to set the
proper roothub port number. Since we've split the xHCI host into two
roothubs, we can't just use the port number in the ancestor hub. Instead,
we loop through the array of hardware port status register speeds and find
the Nth port with a similar speed.
The port status change event handler is updated to figure out whether the
port that reported the change is a USB 3.0 port, or a non-SuperSpeed port.
Once it figures out the port speed, it kicks the proper roothub.
The function to find a slot ID based on the port index is updated to take
into account that the two roothubs will have over-lapping port indexes.
It checks that the virtual device with a matching port index is the same
speed as the passed in roothub.
There's also changes to the driver initialization and shutdown paths:
1. Make sure that the xhci_hcd pointer is shared across the two
usb_hcd structures. The xhci_hcd pointer is allocated and the
registers are mapped in when xhci_pci_setup() is called with the
primary HCD. When xhci_pci_setup() is called with the non-primary
HCD, the xhci_hcd pointer is stored.
2. Make sure to set the sg_tablesize for both usb_hcd structures. Set
the PCI DMA mask for the non-primary HCD to allow for 64-bit or 32-bit
DMA. (The PCI DMA mask is set from the primary HCD further down in
the xhci_pci_setup() function.)
3. Ensure that the host controller doesn't start kicking khubd in
response to port status changes before both usb_hcd structures are
registered. xhci_run() only starts the xHC running once it has been
called with the non-primary roothub. Similarly, the xhci_stop()
function only halts the host controller when it is called with the
non-primary HCD. Then on the second call, it resets and cleans up the
MSI-X irqs.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2010-12-17 02:21:10 +07:00
|
|
|
|
xhci_disable_port(hcd, xhci, wIndex,
|
2010-12-02 02:34:59 +07:00
|
|
|
|
port_array[wIndex], temp);
|
2009-12-10 06:59:11 +07:00
|
|
|
|
break;
|
2012-09-05 12:44:35 +07:00
|
|
|
|
case USB_PORT_FEAT_POWER:
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp & ~PORT_POWER, port_array[wIndex]);
|
2012-09-05 12:44:36 +07:00
|
|
|
|
|
2012-10-15 14:38:34 +07:00
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
2012-09-05 12:44:36 +07:00
|
|
|
|
temp = usb_acpi_power_manageable(hcd->self.root_hub,
|
|
|
|
|
wIndex);
|
|
|
|
|
if (temp)
|
|
|
|
|
usb_acpi_set_power_state(hcd->self.root_hub,
|
|
|
|
|
wIndex, false);
|
2012-10-15 14:38:34 +07:00
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
2012-09-05 12:44:35 +07:00
|
|
|
|
break;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
default:
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
error:
|
|
|
|
|
/* "stall" on error */
|
|
|
|
|
retval = -EPIPE;
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
return retval;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Returns 0 if the status hasn't changed, or the number of bytes in buf.
|
|
|
|
|
* Ports are 0-indexed from the HCD point of view,
|
|
|
|
|
* and 1-indexed from the USB core pointer of view.
|
|
|
|
|
*
|
|
|
|
|
* Note that the status change bits will be cleared as soon as a port status
|
|
|
|
|
* change event is generated, so we use the saved status from that event.
|
|
|
|
|
*/
|
|
|
|
|
int xhci_hub_status_data(struct usb_hcd *hcd, char *buf)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
u32 temp, status;
|
2010-10-14 21:23:00 +07:00
|
|
|
|
u32 mask;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
int i, retval;
|
|
|
|
|
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
|
2011-05-03 20:11:00 +07:00
|
|
|
|
int max_ports;
|
2011-03-29 09:40:46 +07:00
|
|
|
|
__le32 __iomem **port_array;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
struct xhci_bus_state *bus_state;
|
xhci: Avoid "dead ports", add roothub port polling.
The USB core hub thread (khubd) is designed with external USB hubs in
mind. It expects that if a port status change bit is set, the hub will
continue to send a notification through the hub status data transfer.
Basically, it expects hub notifications to be level-triggered.
The xHCI host controller is designed to be edge-triggered on the logical
'OR' of all the port status change bits. When all port status change
bits are clear, and a new change bit is set, the xHC will generate a
Port Status Change Event. If another change bit is set in the same port
status register before the first bit is cleared, it will not send
another event.
This means that the hub code may lose port status changes because of
race conditions between clearing change bits. The user sees this as a
"dead port" that doesn't react to device connects.
The fix is to turn on port polling whenever a new change bit is set.
Once the USB core issues a hub status request that shows that no change
bits are set in any USB ports, turn off port polling.
We can't allow the USB core to poll the roothub for port events during
host suspend because if the PCI host is in D3cold, the port registers
will be all f's. Instead, stop the port polling timer, and
unconditionally restart it when the host resumes. If there are no port
change bits set after the resume, the first call to hub_status_data will
disable polling.
This patch should be backported to stable kernels with the first xHCI
support, 2.6.31 and newer, that include the commit
0f2a79300a1471cf92ab43af165ea13555c8b0a5 "USB: xhci: Root hub support."
There will be merge conflicts because the check for HC_STATE_SUSPENDED
was moved into xhci_suspend in 3.8.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Cc: stable@vger.kernel.org
2012-11-28 03:30:23 +07:00
|
|
|
|
bool reset_change = false;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
2011-05-03 20:11:00 +07:00
|
|
|
|
max_ports = xhci_get_ports(hcd, &port_array);
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state = &xhci->bus_state[hcd_index(hcd)];
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
|
|
|
|
/* Initial status is no changes */
|
2011-05-03 20:11:00 +07:00
|
|
|
|
retval = (max_ports + 8) / 8;
|
2010-05-13 00:20:34 +07:00
|
|
|
|
memset(buf, 0, retval);
|
2012-04-14 01:54:30 +07:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Inform the usbcore about resume-in-progress by returning
|
|
|
|
|
* a non-zero value even if there are no status changes.
|
|
|
|
|
*/
|
|
|
|
|
status = bus_state->resuming_ports;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
|
2015-03-23 23:27:41 +07:00
|
|
|
|
mask = PORT_CSC | PORT_PEC | PORT_OCC | PORT_PLC | PORT_WRC | PORT_CEC;
|
2010-10-14 21:23:00 +07:00
|
|
|
|
|
2009-04-28 09:57:12 +07:00
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
/* For each port, did anything change? If so, set that bit in buf. */
|
2011-05-03 20:11:00 +07:00
|
|
|
|
for (i = 0; i < max_ports; i++) {
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[i]);
|
2010-10-30 04:37:23 +07:00
|
|
|
|
if (temp == 0xffffffff) {
|
|
|
|
|
retval = -ENODEV;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2010-10-14 21:23:00 +07:00
|
|
|
|
if ((temp & mask) != 0 ||
|
2010-12-16 03:47:14 +07:00
|
|
|
|
(bus_state->port_c_suspend & 1 << i) ||
|
|
|
|
|
(bus_state->resume_done[i] && time_after_eq(
|
|
|
|
|
jiffies, bus_state->resume_done[i]))) {
|
2010-05-13 00:20:34 +07:00
|
|
|
|
buf[(i + 1) / 8] |= 1 << (i + 1) % 8;
|
2009-04-28 09:57:12 +07:00
|
|
|
|
status = 1;
|
|
|
|
|
}
|
xhci: Avoid "dead ports", add roothub port polling.
The USB core hub thread (khubd) is designed with external USB hubs in
mind. It expects that if a port status change bit is set, the hub will
continue to send a notification through the hub status data transfer.
Basically, it expects hub notifications to be level-triggered.
The xHCI host controller is designed to be edge-triggered on the logical
'OR' of all the port status change bits. When all port status change
bits are clear, and a new change bit is set, the xHC will generate a
Port Status Change Event. If another change bit is set in the same port
status register before the first bit is cleared, it will not send
another event.
This means that the hub code may lose port status changes because of
race conditions between clearing change bits. The user sees this as a
"dead port" that doesn't react to device connects.
The fix is to turn on port polling whenever a new change bit is set.
Once the USB core issues a hub status request that shows that no change
bits are set in any USB ports, turn off port polling.
We can't allow the USB core to poll the roothub for port events during
host suspend because if the PCI host is in D3cold, the port registers
will be all f's. Instead, stop the port polling timer, and
unconditionally restart it when the host resumes. If there are no port
change bits set after the resume, the first call to hub_status_data will
disable polling.
This patch should be backported to stable kernels with the first xHCI
support, 2.6.31 and newer, that include the commit
0f2a79300a1471cf92ab43af165ea13555c8b0a5 "USB: xhci: Root hub support."
There will be merge conflicts because the check for HC_STATE_SUSPENDED
was moved into xhci_suspend in 3.8.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Cc: stable@vger.kernel.org
2012-11-28 03:30:23 +07:00
|
|
|
|
if ((temp & PORT_RC))
|
|
|
|
|
reset_change = true;
|
|
|
|
|
}
|
|
|
|
|
if (!status && !reset_change) {
|
|
|
|
|
xhci_dbg(xhci, "%s: stopping port polling.\n", __func__);
|
|
|
|
|
clear_bit(HCD_FLAG_POLL_RH, &hcd->flags);
|
2009-04-28 09:57:12 +07:00
|
|
|
|
}
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
return status ? retval : 0;
|
|
|
|
|
}
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_PM
|
|
|
|
|
|
|
|
|
|
int xhci_bus_suspend(struct usb_hcd *hcd)
|
|
|
|
|
{
|
|
|
|
|
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
|
2010-12-16 02:56:29 +07:00
|
|
|
|
int max_ports, port_index;
|
2011-03-29 09:40:46 +07:00
|
|
|
|
__le32 __iomem **port_array;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
struct xhci_bus_state *bus_state;
|
2010-10-14 21:23:03 +07:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
2011-05-03 20:11:00 +07:00
|
|
|
|
max_ports = xhci_get_ports(hcd, &port_array);
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state = &xhci->bus_state[hcd_index(hcd)];
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
|
|
|
|
if (hcd->self.root_hub->do_remote_wakeup) {
|
2012-04-14 01:54:30 +07:00
|
|
|
|
if (bus_state->resuming_ports) {
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
xhci_dbg(xhci, "suspend failed because "
|
|
|
|
|
"a port is resuming\n");
|
|
|
|
|
return -EBUSY;
|
2010-10-14 21:23:03 +07:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2010-12-16 02:56:29 +07:00
|
|
|
|
port_index = max_ports;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state->bus_suspended = 0;
|
2010-12-16 02:56:29 +07:00
|
|
|
|
while (port_index--) {
|
2010-10-14 21:23:03 +07:00
|
|
|
|
/* suspend the port if the port is not suspended */
|
|
|
|
|
u32 t1, t2;
|
|
|
|
|
int slot_id;
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
t1 = readl(port_array[port_index]);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
t2 = xhci_port_state_to_neutral(t1);
|
|
|
|
|
|
|
|
|
|
if ((t1 & PORT_PE) && !(t1 & PORT_PLS_MASK)) {
|
2010-12-16 02:56:29 +07:00
|
|
|
|
xhci_dbg(xhci, "port %d not suspended\n", port_index);
|
2010-12-17 01:49:09 +07:00
|
|
|
|
slot_id = xhci_find_slot_id_by_port(hcd, xhci,
|
2010-12-16 02:56:29 +07:00
|
|
|
|
port_index + 1);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
if (slot_id) {
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
xhci_stop_device(xhci, slot_id, 1);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
}
|
|
|
|
|
t2 &= ~PORT_PLS_MASK;
|
|
|
|
|
t2 |= PORT_LINK_STROBE | XDEV_U3;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
set_bit(port_index, &bus_state->bus_suspended);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
}
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
/* USB core sets remote wake mask for USB 3.0 hubs,
|
2014-11-30 05:47:05 +07:00
|
|
|
|
* including the USB 3.0 roothub, but only if CONFIG_PM
|
USB/xHCI: Enable USB 3.0 hub remote wakeup.
USB 3.0 hubs have a different remote wakeup policy than USB 2.0 hubs.
USB 2.0 hubs, once they have remote wakeup enabled, will always send
remote wakes when anything changes on a port.
However, USB 3.0 hubs have a per-port remote wake up policy that is off
by default. The Set Feature remote wake mask can be changed for any
port, enabling remote wakeup for a connect, disconnect, or overcurrent
event, much like EHCI and xHCI host controller "wake on" port status
bits. The bits are cleared to zero on the initial hub power on, or
after the hub has been reset.
Without this patch, when a USB 3.0 hub gets suspended, it will not send
a remote wakeup on device connect or disconnect. This would show up to
the user as "dead ports" unless they ran lsusb -v (since newer versions
of lsusb use the sysfs files, rather than sending control transfers).
Change the hub driver's suspend method to enable remote wake up for
disconnect, connect, and overcurrent for all ports on the hub. Modify
the xHCI driver's roothub code to handle that request, and set the "wake
on" bits in the port status registers accordingly.
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
2012-01-07 01:34:31 +07:00
|
|
|
|
* is enabled, so also enable remote wake here.
|
|
|
|
|
*/
|
2014-11-18 16:27:13 +07:00
|
|
|
|
if (hcd->self.root_hub->do_remote_wakeup) {
|
2010-10-14 21:23:03 +07:00
|
|
|
|
if (t1 & PORT_CONNECT) {
|
|
|
|
|
t2 |= PORT_WKOC_E | PORT_WKDISC_E;
|
|
|
|
|
t2 &= ~PORT_WKCONN_E;
|
|
|
|
|
} else {
|
|
|
|
|
t2 |= PORT_WKOC_E | PORT_WKCONN_E;
|
|
|
|
|
t2 &= ~PORT_WKDISC_E;
|
|
|
|
|
}
|
|
|
|
|
} else
|
|
|
|
|
t2 &= ~PORT_WAKE_BITS;
|
|
|
|
|
|
|
|
|
|
t1 = xhci_port_state_to_neutral(t1);
|
|
|
|
|
if (t1 != t2)
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(t2, port_array[port_index]);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
}
|
|
|
|
|
hcd->state = HC_STATE_SUSPENDED;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state->next_statechange = jiffies + msecs_to_jiffies(10);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int xhci_bus_resume(struct usb_hcd *hcd)
|
|
|
|
|
{
|
|
|
|
|
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
|
2010-12-16 02:56:29 +07:00
|
|
|
|
int max_ports, port_index;
|
2011-03-29 09:40:46 +07:00
|
|
|
|
__le32 __iomem **port_array;
|
2010-12-16 03:47:14 +07:00
|
|
|
|
struct xhci_bus_state *bus_state;
|
2010-10-14 21:23:03 +07:00
|
|
|
|
u32 temp;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
2011-05-03 20:11:00 +07:00
|
|
|
|
max_ports = xhci_get_ports(hcd, &port_array);
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state = &xhci->bus_state[hcd_index(hcd)];
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
2010-12-16 03:47:14 +07:00
|
|
|
|
if (time_before(jiffies, bus_state->next_statechange))
|
2010-10-14 21:23:03 +07:00
|
|
|
|
msleep(5);
|
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
if (!HCD_HW_ACCESSIBLE(hcd)) {
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
return -ESHUTDOWN;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* delay the irqs */
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(&xhci->op_regs->command);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
temp &= ~CMD_EIE;
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, &xhci->op_regs->command);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
2010-12-16 02:56:29 +07:00
|
|
|
|
port_index = max_ports;
|
|
|
|
|
while (port_index--) {
|
2010-10-14 21:23:03 +07:00
|
|
|
|
/* Check whether need resume ports. If needed
|
|
|
|
|
resume port and disable remote wakeup */
|
|
|
|
|
u32 temp;
|
|
|
|
|
int slot_id;
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(port_array[port_index]);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
if (DEV_SUPERSPEED(temp))
|
|
|
|
|
temp &= ~(PORT_RWC_BITS | PORT_CEC | PORT_WAKE_BITS);
|
|
|
|
|
else
|
|
|
|
|
temp &= ~(PORT_RWC_BITS | PORT_WAKE_BITS);
|
2010-12-16 03:47:14 +07:00
|
|
|
|
if (test_bit(port_index, &bus_state->bus_suspended) &&
|
2010-10-14 21:23:03 +07:00
|
|
|
|
(temp & PORT_PLS_MASK)) {
|
|
|
|
|
if (DEV_SUPERSPEED(temp)) {
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array,
|
|
|
|
|
port_index, XDEV_U0);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
} else {
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array,
|
|
|
|
|
port_index, XDEV_RESUME);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
msleep(20);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
2011-09-24 04:19:48 +07:00
|
|
|
|
xhci_set_link_state(xhci, port_array,
|
|
|
|
|
port_index, XDEV_U0);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
}
|
2011-04-19 16:17:39 +07:00
|
|
|
|
/* wait for the port to enter U0 and report port link
|
|
|
|
|
* state change.
|
|
|
|
|
*/
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
msleep(20);
|
|
|
|
|
spin_lock_irqsave(&xhci->lock, flags);
|
|
|
|
|
|
|
|
|
|
/* Clear PLC */
|
2011-09-24 04:19:49 +07:00
|
|
|
|
xhci_test_and_clear_bit(xhci, port_array, port_index,
|
|
|
|
|
PORT_PLC);
|
2011-04-19 16:17:39 +07:00
|
|
|
|
|
2010-12-17 01:49:09 +07:00
|
|
|
|
slot_id = xhci_find_slot_id_by_port(hcd,
|
|
|
|
|
xhci, port_index + 1);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
if (slot_id)
|
|
|
|
|
xhci_ring_device(xhci, slot_id);
|
|
|
|
|
} else
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, port_array[port_index]);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
}
|
|
|
|
|
|
2013-11-15 10:34:06 +07:00
|
|
|
|
(void) readl(&xhci->op_regs->command);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
2010-12-16 03:47:14 +07:00
|
|
|
|
bus_state->next_statechange = jiffies + msecs_to_jiffies(5);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
/* re-enable irqs */
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(&xhci->op_regs->command);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
temp |= CMD_EIE;
|
2013-11-15 10:34:07 +07:00
|
|
|
|
writel(temp, &xhci->op_regs->command);
|
2013-11-15 10:34:06 +07:00
|
|
|
|
temp = readl(&xhci->op_regs->command);
|
2010-10-14 21:23:03 +07:00
|
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xhci->lock, flags);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2010-10-16 04:59:15 +07:00
|
|
|
|
#endif /* CONFIG_PM */
|