The function dma_set_mask() tests internally whether the dma_mask pointer
for the device is initialized and fails if the dma_mask pointer is NULL.
On pci platforms, the device dma_mask pointer is initialized, when pci
devices are enumerated, to point to the pci_dev->dma_mask which is 0xffffffff.
However, for non-pci platforms, the dma_mask pointer may not be initialized
and in that case dma_set_mask() will fail.

This patch initializes the dma_mask and the coherent_dma_mask to 32bits
in xhci_plat_probe(), before the call to usb_create_hcd() that sets the
"uses_dma" flag for the usb bus and the call to usb_add_hcd() that creates
coherent dma pools for the usb hcd.

Moreover, a call to dma_set_mask() does not set the device coherent_dma_mask.
Since the xhci-hcd driver calls dma_alloc_coherent() and dma_pool_alloc()
to allocate consistent DMA memory blocks, the coherent DMA address mask
has to be set explicitly.

This patch sets the coherent_dma_mask to 64bits in xhci_gen_setup() when
the xHC is capable for 64-bit DMA addressing.

If dma_set_mask() succeeds, for a given bitmask, it is guaranteed that
the given bitmask is also supported for consistent DMA mappings.

Other changes introduced in this patch are:

- The return value of dma_set_mask() is checked to ensure that the required
  dma bitmask conforms with the host system's addressing capabilities.

- The dma_mask setup code for the non-primary hcd was removed since both
  primary and non-primary hcd refer to the same generic device whose
  dma_mask and coherent_dma_mask are already set during the setup of
  the primary hcd.

- The code for reading the HCCPARAMS register to find out the addressing
  capabilities of xHC was removed since its value is already cached in
  xhci->hccparams.

- hcd->self.controller was replaced with the dev variable since it is
  already available.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new trace event, which is called xhci_dbg_ring_expansion
and belongs to the event class xhci_log_msg, and adds tracepoints that trace
the debug messages associated with the expansion of endpoint ring when there
is not enough space allocated to hold all pending TRBs.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new trace event, which is called xhci_dbg_init
and belongs to the event class xhci_log_msg, and adds tracepoints that
trace the debug statements in the functions used to start and stop the
xhci-hcd driver.

Also, it removes an unnecessary cast of variable val to unsigned int
in xhci_mem_init(), since val is already declared as unsigned int.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new trace event, which is called xhci_dbg_cancel_urb
and belongs to the event class xhci_log_msg, and adds tracepoints that
trace the debug messages related to the removal of a cancelled URB from
the endpoint's transfer ring.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch creates a new event class, called xhci_log_event,
and defines the xhci_cmd_completion trace event used for
tracing the commands issued to xHC that generate a completion
event in the event ring.

This info can be used, later, to print, in a human readable
way, the completion status and flags as well as the command's
type and fields using the trace-cmd tool and the appropriate
plugin.

Also, a tracepoint is added in handle_cmd_completion().
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new event class, called xhci_log_ctx,
that records in the ring buffer the context data, the
context type (input or output), the context dma and virtual
addresses, the context endpoint entries, the slot ID and
whether the xHC uses 64 byte context data structures.

This information can be used, later, to parse and display
the context data fields with the appropriate plugin using
the trace-cmd tool.

Also, this patch defines a trace event, called xhci_address_ctx,
to trace the contexts related to the Address Device command and
adds the associated tracepoints in xhci_address_device().
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new trace event, which is called xhci_dbg_reset_ep
and belongs in the event class xhci_log_msg, and adds tracepoints that
trace the debug messages associated with resetting an endpoint after
the reception of a STALL packet.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new trace event, which is called xhci_dbg_quirks
and belongs in the event class xhci_log_msg, and adds tracepoints that
trace the debug messages associated with xHCs' quirks.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch defines a new trace event, which is called xhci_dbg_context_change
and belongs in the event class xhci_log_msg, and adds tracepoints for tracing
the debug messages related to context updates performed with Configure Endpoint
and Evaluate Context commands.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch declares an event class for trace events that
trace messages with variadic arguments, called xhci_log_msg,
and defines a trace event for tracing the debug messages in
xhci_address_device() function, called xhci_dbg_address.

In order to implement this type of trace events, a wrapper function,
called xhci_dbg_trace(), was created that records the format string
and variadic arguments into a va_format structure which is passed as
argument to the tracepoints of the class xhci_log_msg.

All the xhci_dbg() calls in xhci_address_device() are replaced
with calls to xhci_dbg_trace(). The functionality of xhci_dbg()
log messages was not removed though, but it is placed inside
xhci_dbg_trace().

This trace event aims to give the ability to the user or the
developper to isolate and trace the debug messages generated
when an Address Device Command is issued to xHC.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

CONFIG_USB_XHCI_HCD_DEBUGGING option is used to enable
verbose debugging output for the xHCI host controller
driver.

In the current version of the xhci-hcd driver, this
option must be turned on, in order for the debugging
log messages to be displayed, and users may need to
recompile the linux kernel to obtain debugging
information that will help them track down problems.

This patch removes the above debug option to enable
debugging log messages at all times.
The aim of this is to rely on the debugfs and the
dynamic debugging feature for fine-grained management
of debugging messages and to not force users to set
the debug config option and compile the linux kernel
in order to have access in that information.

This patch, also, removes the XHCI_DEBUG symbol and the
functions dma_to_stream_ring(), xhci_test_radix_tree()
and xhci_event_ring_work() that are not useful anymore.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch replaces the calls to printk(KERN_DEBUG ...)
with either calls to xhci_dbg() or calls to pr_debug(),
depending on whether the xhci_hcd structure is available
at callsite, so that the correspoding debugging messages
are not enabled by default when CONFIG_DYNAMIC_DEBUG option
is set but rather can be enabled dynamically taking advantage
of the dynamic debugging feature.

Also, it adds a newline at the end of debugging messages in
case there is not, so that messages don't appear broken
when printed.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

This patch replaces the calls to xhci_info() with calls to
xhci_dbg() and removes the unused xhci_info() definition
from xhci-hcd.

By replacing the xhci_info() with xhci_dbg(), the calls to
dev_info() are replaced with calls to dev_dbg() so that
their output can be dynamically controlled via the dynamic
debugging mechanism.
Signed-off-by: NXenia Ragiadakou <burzalodowa@gmail.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

Add Device Tree match table to xhci-plat.c. Add DT bindings document.
Signed-off-by: NAl Cooper <alcooperx@gmail.com>
Cc: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Cc: Felipe Balbi <balbi@ti.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

The AT91 PMC (Power Management Controller) provides an USB clock used by
USB Full Speed host (ohci) and USB Full Speed device (udc).
The usb drivers (ohci and udc) must configure this clock to 48Mhz.
This configuration was formely done in mach-at91/clock.c, but this
implementation will be removed when moving to common clk framework.

This patch adds support for usb clock retrieval and configuration, and is
backward compatible with the current at91 clk implementation (if usb clk
is not found, it does not configure/enable it).

Changes since v1:
 - use IS_ENABLED(CONFIG_COMMON_CLK) to isolate new at91 clk support
Signed-off-by: NBoris BREZILLON <b.brezillon@overkiz.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In theory, an EHCI controller can turn off the PORT_RESUME or
PORT_RESET bits in a port status register all by itself (and some
controllers actually do this).  We shouldn't depend on these bits
being set correctly.

This patch rearranges the code in ehci-hcd that handles completion of
port resets and resumes.  We guarantee that ehci->reset_done[portnum]
is nonzero if a reset or resume is in progress, and that the portnum
bit is set in ehci->resuming_ports if the operation is a resume.  (To
help enforce this guarantee, the patch prevents suspended ports from
being reset.)  Therefore it's not necessary to look at the port status
bits to learn what's going on.

The patch looks bigger than it really is, because it changes the
indentation level of a sizeable region of code.  Most of what it
actually does is interchange some tests.  The only functional changes
are testing reset_done and resuming_ports rather than PORT_RESUME and
PORT_RESET, removing a now-unnecessary check for spontaneous
resets of the PORT_RESUME and PORT_RESET bits, and preventing a
suspended or resuming port from being reset.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The ehci-hcd driver isn't as careful as it should be about the way it
uses ehci->resuming_ports.  One of the omissions was fixed recently by
commit 47a64a13 (USB: EHCI: Fix resume signalling on remote
wakeup), but there are other places that need attention:

	When a port's suspend feature is explicitly cleared, the
	corresponding bit in resuming_ports should be set and the core
	should be notified about the port resume.

	We don't need to clear a resuming_ports bit when a reset
	completes.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

FOTG210 is an OTG controller which can be configured as an
USB2.0 host. FOTG210 host is an ehci-like controller with
some differences. First, register layout of FOTG210 is
incompatible with EHCI. Furthermore, FOTG210 is lack of
siTDs which means iTDs are used for both HS and FS ISO
transfer.
Signed-off-by: NFeng-Hsin Chiang <john453@faraday-tech.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When ohci-hcd is shutting down, call ohci_usb_reset reset ohci-hcd, the
root hub generate an interrupt, but ohci->rh_state is OHCI_RH_HALTED,
and ohci_irq ignore the interrupt, the kernel trigger warning "irq
nobody cared". ehci-hcd is first disable interrupts, then reset ehci.

This patch disable ohci interrupt before reset ohci.

The patch is tested at the arm cortex-a9 demo board.
Signed-off-by: Ncaizhiyong <caizhiyong@huawei.com>
Reviewed-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge the usb_hcd_ep93xx_probe() into ohci_hcd_ep93xx_drv_probe() and
the usb_hcd_ep93xx_remove() into ohci_hcd_ep93xx_drv_remove(). As Alan
Stern pointed out, there is no reason for them to be separate.

Also, as Alan Stern suggested, eliminate the ep93xx_start_hc() and
ep93xx_stop_hc() routines and simply call clk_enable() and clk_disable()
directly. The extra level of redirection does not add any clarity.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Cc: Lennert Buytenhek <kernel@wantstofly.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use devm_clk_get() to make the code a bit cleaner and simpler.

This also fixes a bug where a clk_put() is not done if usb_add_hcd()
fails.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Cc: Lennert Buytenhek <kernel@wantstofly.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use platform_get_irq() instead of accessing the platform_device
resources directly.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Cc: Lennert Buytenhek <kernel@wantstofly.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use devm_ioremap_resource() to make the code a bit cleaner and
simpler.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Cc: Lennert Buytenhek <kernel@wantstofly.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Most HCD drivers are doing the same thing in their ".shutdown" callback
so it makes sense to use the generic usb_hcd_platform_shutdown()
handler there.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NFelipe Balbi <balbi@ti.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

in some cases where device is attched to xhci port and do not responding,
for example ath9k_htc with stalled firmware, kernel will
crash on ring_doorbell_for_active_rings.
This patch check if pointer exist before it is used.

This patch should be backported to kernels as old as 2.6.35, that
contain the commit e9df17eb "USB: xhci:
Correct assumptions about number of rings per endpoint"
Signed-off-by: NOleksij Rempel <linux@rempel-privat.de>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org

Xhci controllers with hci_version > 0.96 gives spurious success
events on short packet completion. During webcam capture the
"ERROR Transfer event TRB DMA ptr not part of current TD" was observed.
The same application works fine with synopsis controllers hci_version 0.96.
The same issue is seen with Intel Pantherpoint xhci controller. So enabling
this quirk in xhci_gen_setup if controller verion is greater than 0.96.
For xhci-pci move the quirk to much generic place xhci_gen_setup.

Note from Sarah:

The xHCI 1.0 spec changed how hardware handles short packets.  The HW
will notify SW of the TRB where the short packet occurred, and it will
also give a successful status for the last TRB in a TD (the one with the
IOC flag set).  On the second successful status, that warning will be
triggered in the driver.

Software is now supposed to not assume the TD is not completed until it
gets that last successful status.  That means we have a slight race
condition, although it should have little practical impact.  This patch
papers over that issue.

It's on my long-term to-do list to fix this race condition, but it is a
much more involved patch that will probably be too big for stable.  This
patch is needed for stable to avoid serious log spam.

This patch should be backported to kernels as old as 3.0, that
contain the commit ad808333 "Intel xhci:
Ignore spurious successful event."

The patch will have to be modified for kernels older than 3.2, since
that kernel added the xhci_gen_setup function for xhci platform devices.
The correct conflict resolution for kernels older than 3.2 is to set
XHCI_SPURIOUS_SUCCESS in xhci_pci_quirks for all xHCI 1.0 hosts.
Signed-off-by: NGeorge Cherian <george.cherian@ti.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@vger.kernel.org

Fix warning when CONFIG_PCI is not enabled
(from commit 29636578).

drivers/usb/host/pci-quirks.h: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Reported-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Cc: Moiz Sonasath <m-sonasath@ti.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

Resolves the following build warnings:
drivers/usb/host/xhci.c:332:13: warning: 'xhci_msix_sync_irqs' defined but not used [-Wunused-function]
drivers/usb/host/xhci.c:3901:12: warning: 'xhci_change_max_exit_latency' defined but not used [-Wunused-function]

These functions are not always used, and since they're marked static
they will produce build warnings:
- xhci_msix_sync_irqs is only used with CONFIG_PCI.
- xhci_change_max_exit_latency is a little more complicated with
  dependencies on CONFIG_PM and CONFIG_PM_RUNTIME.

Instead of building a bigger maze of ifdefs in this code, I've just
marked both with __maybe_unused.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

When the host controller fails to respond to an Enable Slot command, and
the host fails to respond to the register write to abort the command
ring, the xHCI driver will assume the host is dead, and call
usb_hc_died().

The USB device's slot_id is still set to zero, and the pointer stored at
xhci->devs[0] will always be NULL.  The call to xhci_check_args in
xhci_free_dev should have caught the NULL virt_dev pointer.

However, xhci_free_dev is designed to free the xhci_virt_device
structures, even if the host is dead, so that we don't leak kernel
memory.  xhci_free_dev checks the return value from the generic
xhci_check_args function.  If the return value is -ENODEV, it carries on
trying to free the virtual device.

The issue is that xhci_check_args looks at the host controller state
before it looks at the xhci_virt_device pointer.  It will return -ENIVAL
because the host is dead, and xhci_free_dev will ignore the return
value, and happily dereference the NULL xhci_virt_device pointer.

The fix is to make sure that xhci_check_args checks the xhci_virt_device
pointer before it checks the host state.

See https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1203453 for
further details.  This patch doesn't solve the underlying issue, but
will ensure we don't see any more NULL pointer dereferences because of
the issue.

This patch should be backported to kernels as old as 3.1, that
contain the commit 7bd89b40 "xhci: Don't
submit commands or URBs to halted hosts."
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Reported-by: NVincent Thiele <vincentthiele@gmail.com>
Cc: stable@vger.kernel.org

This reverts commit 1dd3d123.

The email address for the developer now bounces, which means they have
moved on, so remove the driver until someone else from the company steps
up to maintain it.
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

FOTG210 is an OTG controller which can be configured as an
USB2.0 host. FOTG210 host is an ehci-like controller with
some differences. First, register layout of FOTG210 is
incompatible with EHCI. Furthermore, FOTG210 is lack of
siTDs which means iTDs are used for both HS and FS ISO
transfer.
Signed-off-by: NYuan-Hsin Chen <yhchen@faraday-tech.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Drivers should not be putting debug files in /proc/ that is what debugfs
is for, so move the isp1362 driver's debug file to debugfs.
Reviewed-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Drivers should not be putting debug files in /proc/ that is what debugfs
is for, so move the sl811 driver's debug file to debugfs.
Reviewed-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This removes the dependency of the driver on CONFIG_USB_DEBUG and moves
it to us the dynamic debug subsystem instead.  Bonus is the fact that we
can now properly determine the exact hardware that is spitting out the
messages.

This lets debugging be enabled without having to rebuild the driver, an
important thing for users that can not do it.
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move all debugging messages in the driver to use the dynamic debug
subsystem, and not rely on CONFIG_USB_DEBUG to turn them on or off.

This lets debugging be enabled without having to rebuild the driver, an
important thing for users that can not do it.

It also removes the pointless IRQ_TEST() macro, as that was totally
useless and obviously never used.

Cc: Olav Kongas <ok@artecdesign.ee>
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Now that the debugging macros are cleaned up, just rely on the dynamic
debug code in the kernel to do the debug messages for the driver.

This lets debugging be enabled without having to rebuild the driver, an
important thing for users that can not do it.
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

If you want a debug call, just make it, so move to using the
already-there DBG() call.  No need to make things more complex than they
really need to be.
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Like _BUG_ON(), _WARN_ON() wasn't ever being used, so just delete it, as
obviously things are working properly now (if not, we have bigger
problems...)
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

We shouldn't ever panic in a driver, and these calls were never being
used, so just delete them, as obviously the driver is working properly
now (right?)
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>