openwrt/staging/blogic.git
5 years agousb: musb: jz4740: Let the platform probe the PHY
Paul Cercueil [Tue, 30 Apr 2019 14:59:39 +0000 (09:59 -0500)]
usb: musb: jz4740: Let the platform probe the PHY

By registering a generic USB PHY from within the driver, we may shadow
the USB PHY registered by the platform, which might be different.

Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Bin Liu <b-liu@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb: Add usb-phy property to the jz4740-musb node
Paul Cercueil [Tue, 30 Apr 2019 14:59:38 +0000 (09:59 -0500)]
dt-bindings: usb: Add usb-phy property to the jz4740-musb node

Add a required 'usb-phy' property, to obtain a phandle to the USB PHY
from devicetree.

Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Bin Liu <b-liu@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agosoc: sunxi: Fix missing dependency on REGMAP_MMIO
Samuel Holland [Tue, 30 Apr 2019 14:59:37 +0000 (09:59 -0500)]
soc: sunxi: Fix missing dependency on REGMAP_MMIO

When enabling ARCH_SUNXI from allnoconfig, SUNXI_SRAM is enabled, but
not REGMAP_MMIO, so the kernel fails to link with an undefined reference
to __devm_regmap_init_mmio_clk. Select REGMAP_MMIO, as suggested in
drivers/base/regmap/Kconfig.

This creates the following dependency loop:

  drivers/of/Kconfig:68:                symbol OF_IRQ depends on IRQ_DOMAIN
  kernel/irq/Kconfig:63:                symbol IRQ_DOMAIN is selected by REGMAP
  drivers/base/regmap/Kconfig:7:        symbol REGMAP default is visible depending on REGMAP_MMIO
  drivers/base/regmap/Kconfig:39:       symbol REGMAP_MMIO is selected by SUNXI_SRAM
  drivers/soc/sunxi/Kconfig:4:          symbol SUNXI_SRAM is selected by USB_MUSB_SUNXI
  drivers/usb/musb/Kconfig:63:          symbol USB_MUSB_SUNXI depends on GENERIC_PHY
  drivers/phy/Kconfig:7:                symbol GENERIC_PHY is selected by PHY_BCM_NS_USB3
  drivers/phy/broadcom/Kconfig:29:      symbol PHY_BCM_NS_USB3 depends on MDIO_BUS
  drivers/net/phy/Kconfig:12:           symbol MDIO_BUS default is visible depending on PHYLIB
  drivers/net/phy/Kconfig:181:          symbol PHYLIB is selected by ARC_EMAC_CORE
  drivers/net/ethernet/arc/Kconfig:18:  symbol ARC_EMAC_CORE is selected by ARC_EMAC
  drivers/net/ethernet/arc/Kconfig:24:  symbol ARC_EMAC depends on OF_IRQ

To fix the circular dependency, make USB_MUSB_SUNXI select GENERIC_PHY
instead of depending on it. This matches the use of GENERIC_PHY by all
but two other drivers.

Cc: <stable@vger.kernel.org> # 4.19
Fixes: 5828729bebbb ("soc: sunxi: export a regmap for EMAC clock reg on A64")
Signed-off-by: Samuel Holland <samuel@sholland.org>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Bin Liu <b-liu@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: musb: omap2430: Add support for idling phy when musb is idle
Tony Lindgren [Tue, 30 Apr 2019 14:59:36 +0000 (09:59 -0500)]
usb: musb: omap2430: Add support for idling phy when musb is idle

I noticed that musb is blocking core retention for omap4 unlike for
omap3. This is because for omap3 we have phy-twl4030-usb implement
it's own PM runtime to handle errata "VUSB3V1 VBUS overvoltage
debouncer not working when the PHY is powered down". That is done
in order to keep the USB PHY powered when phy-twl4030-usb is loaded.

For the other USB PHYs, we need to enable and disable the PHY based on
musb PM runtime. With the session bit based PM runtime for musb core,
we can now idle the USB PHY always when musb is idle.

Note that adding these calls will not affect the twl4030 driver
as it's phy functions will just query the PHY state without powering
the PHY on or off.

Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Bin Liu <b-liu@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: musb: Silence error about blacklisting hubs if !CONFIG_USB
Paul Cercueil [Tue, 30 Apr 2019 14:59:35 +0000 (09:59 -0500)]
usb: musb: Silence error about blacklisting hubs if !CONFIG_USB

Some drivers, like jz4740-musb, don't depend on CONFIG_USB.

Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Bin Liu <b-liu@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousbip: vhci_hcd: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 29 Apr 2019 14:39:57 +0000 (09:39 -0500)]
usbip: vhci_hcd: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.

This patch fixes the following warning:

In file included from drivers/usb/usbip/vhci_hcd.c:15:
drivers/usb/usbip/vhci_hcd.c: In function ‘vhci_hub_control’:
drivers/usb/usbip/usbip_common.h:63:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (flag & usbip_debug_flag)  \
      ^
drivers/usb/usbip/usbip_common.h:77:2: note: in expansion of macro ‘usbip_dbg_with_flag’
  usbip_dbg_with_flag(usbip_debug_vhci_rh, fmt , ##args)
  ^~~~~~~~~~~~~~~~~~~
drivers/usb/usbip/vhci_hcd.c:509:4: note: in expansion of macro ‘usbip_dbg_vhci_rh’
    usbip_dbg_vhci_rh(
    ^~~~~~~~~~~~~~~~~
drivers/usb/usbip/vhci_hcd.c:511:3: note: here
   case USB_PORT_FEAT_U2_TIMEOUT:
   ^~~~

Warning level 3 was used: -Wimplicit-fallthrough=3

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoUAS: fix alignment of scatter/gather segments
Oliver Neukum [Tue, 30 Apr 2019 10:21:45 +0000 (12:21 +0200)]
UAS: fix alignment of scatter/gather segments

This is the UAS version of

747668dbc061b3e62bc1982767a3a1f9815fcf0e
usb-storage: Set virt_boundary_mask to avoid SG overflows

We are not as likely to be vulnerable as storage, as it is unlikelier
that UAS is run over a controller without native support for SG,
but the issue exists.
The issue has been existing since the inception of the driver.

Fixes: 115bb1ffa54c ("USB: Add UAS driver")
Signed-off-by: Oliver Neukum <oneukum@suse.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoUSB: cdc-acm: clean up throttle handling
Johan Hovold [Thu, 25 Apr 2019 16:05:40 +0000 (18:05 +0200)]
USB: cdc-acm: clean up throttle handling

Clean up the throttle implementation by dropping the redundant
throttle_req flag which was a remnant from back when USB serial had only
a single read URB, something which was later carried over to cdc-acm.

Also convert the throttled flag to an atomic bit flag.

Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoUSB: cdc-acm: fix unthrottle races
Johan Hovold [Thu, 25 Apr 2019 16:05:39 +0000 (18:05 +0200)]
USB: cdc-acm: fix unthrottle races

Fix two long-standing bugs which could potentially lead to memory
corruption or leave the port throttled until it is reopened (on weakly
ordered systems), respectively, when read-URB completion races with
unthrottle().

First, the URB must not be marked as free before processing is complete
to prevent it from being submitted by unthrottle() on another CPU.

CPU 1 CPU 2
================ ================
complete() unthrottle()
  process_urb();
  smp_mb__before_atomic();
  set_bit(i, free);   if (test_and_clear_bit(i, free))
  submit_urb();

Second, the URB must be marked as free before checking the throttled
flag to prevent unthrottle() on another CPU from failing to observe that
the URB needs to be submitted if complete() sees that the throttled flag
is set.

CPU 1 CPU 2
================ ================
complete() unthrottle()
  set_bit(i, free);   throttled = 0;
  smp_mb__after_atomic();   smp_mb();
  if (throttled)   if (test_and_clear_bit(i, free))
  return;   submit_urb();

Note that test_and_clear_bit() only implies barriers when the test is
successful. To handle the case where the URB is still in use an explicit
barrier needs to be added to unthrottle() for the second race condition.

Also note that the first race was fixed by 36e59e0d70d6 ("cdc-acm: fix
race between callback and unthrottle") back in 2015, but the bug was
reintroduced a year later.

Fixes: 1aba579f3cf5 ("cdc-acm: handle read pipe errors")
Fixes: 088c64f81284 ("USB: cdc-acm: re-write read processing")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Oliver Neukum <oneukum@suse.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: ucsi: ccg: fix missing unlock on error in ccg_cmd_write_flash_row()
Wei Yongjun [Mon, 29 Apr 2019 12:26:30 +0000 (12:26 +0000)]
usb: typec: ucsi: ccg: fix missing unlock on error in ccg_cmd_write_flash_row()

Add the missing unlock before return from function ccg_cmd_write_flash_row()
in the error handling case.

Fixes: 5c9ae5a87573 ("usb: typec: ucsi: ccg: add firmware flashing support")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb: renesas_usbhs: Add support for r8a77470
Biju Das [Mon, 29 Apr 2019 10:22:57 +0000 (11:22 +0100)]
dt-bindings: usb: renesas_usbhs: Add support for r8a77470

Document support for RZ/G1C (R8A77470) SoC.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb-xhci: Add r8a774c0 support
Fabrizio Castro [Mon, 29 Apr 2019 08:51:24 +0000 (09:51 +0100)]
dt-bindings: usb-xhci: Add r8a774c0 support

Document RZ/G2E (R8A774C0) SoC bindings.

Signed-off-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: xhci: add endpoint context tracing when an endpoint is added
Mathias Nyman [Fri, 26 Apr 2019 13:23:32 +0000 (16:23 +0300)]
usb: xhci: add endpoint context tracing when an endpoint is added

The configure endpoint command configures all the endpoints that were
flagged to be added or dropped.

To know the content of each of the added endpoints we need to add tracing
to the .add_endpoint() callback, just after initializing all the context
values.

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoxhci: Add tracing for input control context
Mathias Nyman [Fri, 26 Apr 2019 13:23:31 +0000 (16:23 +0300)]
xhci: Add tracing for input control context

Add tracing for the add and drop bits in the input control context
used in Address device, configure endpoint, evaluate context commands.

The add and drop bits tell xHC which enpoints are added and dropped.

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoxhci: add port and bus number to port dynamic debugging
Mathias Nyman [Fri, 26 Apr 2019 13:23:30 +0000 (16:23 +0300)]
xhci: add port and bus number to port dynamic debugging

Improve port related dynamic debugging by printing out the bus number,
port number and port status register content each time there is a port
related debug messages.

Use the same port numbering method as usbcore to simplify debugging.
i.e. starting with port number 1.

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: xhci: add Immediate Data Transfer support
Nicolas Saenz Julienne [Fri, 26 Apr 2019 13:23:29 +0000 (16:23 +0300)]
usb: xhci: add Immediate Data Transfer support

Immediate data transfers (IDT) allow the HCD to copy small chunks of
data (up to 8bytes) directly into its output transfer TRBs. This avoids
the somewhat expensive DMA mappings that are performed by default on
most URBs submissions.

In the case an URB was suitable for IDT. The data is directly copied
into the "Data Buffer Pointer" region of the TRB and the IDT flag is
set. Instead of triggering memory accesses the HC will use the data
directly.

The implementation could cover all kind of output endpoints. Yet
Isochronous endpoints are bypassed as I was unable to find one that
matched IDT's constraints. As we try to bypass the default DMA mappings
on URB buffers we'd need to find a Isochronous device with an
urb->transfer_buffer_length <= 8 bytes.

The implementation takes into account that the 8 byte buffers provided
by the URB will never cross a 64KB boundary.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Reviewed-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: usb251xb: Add an empty hub' i2c-bus segment checker
Serge Semin [Sat, 27 Apr 2019 09:06:44 +0000 (12:06 +0300)]
usb: usb251xb: Add an empty hub' i2c-bus segment checker

It's pointless to scan the hub' i2c-bus segment if GPIOs aren't supported
by the system, since no GPIO-driven reset could be cleared by the driver
then. Moreover if CONFIG_GPIOLIB is disabled the gpio_chip structure
definition won't be available, which causes the incomplete type pointer
dereference compilation error. In order to fix this we need to create an
empty usb251x_check_gpio_chip() method returning zero, so the driver would
skip the i2c-bus segment checking and proceed with further probing in this
case.

Fixes: 6e3c8beb4f92 ("usb: usb251xb: Lock i2c-bus segment the hub resides")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb/hcd: Send a uevent signaling that the host controller had died
Raul E Rangel [Fri, 19 Apr 2019 15:30:22 +0000 (09:30 -0600)]
usb/hcd: Send a uevent signaling that the host controller had died

This change will send an OFFLINE event to udev with the ERROR=DEAD
environment variable set when the HC dies.

By notifying user space the appropriate policies can be applied.
i.e.,
 * Collect error logs.
 * Notify the user that USB is no longer functional.
 * Perform a graceful reboot.

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: Add driver for NVIDIA Alt Modes
Ajay Gupta [Tue, 23 Apr 2019 14:21:51 +0000 (17:21 +0300)]
usb: typec: Add driver for NVIDIA Alt Modes

Latest NVIDIA GPUs support VirtualLink device. Since USBIF
has not assigned a Standard ID (SID) for VirtualLink
so using NVIDA VID 0x955 as SVID.

Signed-off-by: Ajay Gupta <ajayg@nvidia.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: displayport: Export probe and remove functions
Ajay Gupta [Tue, 23 Apr 2019 14:21:50 +0000 (17:21 +0300)]
usb: typec: displayport: Export probe and remove functions

VirtualLink standard extends the DisplayPort Alt Mode by
utilizing also the USB 2 pins on the USB Type-C connector.
It uses the same messages as DisplayPort, but not the DP
SVID. At the time of writing, USB IF has not assigned a
Standard ID (SID) for VirtualLink, so the manufacturers of
VirtualLink adapters use their Vendor IDs as the SVID.

Since the SVID specific communication is exactly the same as
with DisplayPort alternate mode, there is no need to
implement separate driver for VirtualLink. We'll handle the
current VirtualLink adapters with probe drivers, and once
there is SVID assigned for it, we add it to the displayport
alt mode driver.

To support probing drivers, exporting the probe and remove
functions, and also changing the DP_HEADER helper macro to
use the SVID of the alternate mode device instead of the
DisplayPort alt mode SVID.

Suggested-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Ajay Gupta <ajayg@nvidia.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: ucsi: Support for DisplayPort alt mode
Heikki Krogerus [Tue, 23 Apr 2019 14:21:49 +0000 (17:21 +0300)]
usb: typec: ucsi: Support for DisplayPort alt mode

This makes it possible to bind a driver to a DisplayPort
alt mode adapter devices.

The driver attempts to cope with the limitations of UCSI by
"emulating" behaviour and attempting to guess things when
ever possible in order to satisfy the requirements the
standard DisplayPort alt mode driver has.

Tested-by: Ajay Gupta <ajayg@nvidia.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: ucsi: Preliminary support for alternate modes
Heikki Krogerus [Tue, 23 Apr 2019 14:21:48 +0000 (17:21 +0300)]
usb: typec: ucsi: Preliminary support for alternate modes

With UCSI the alternate modes, just like everything else
related to USB Type-C connectors, are handled in firmware.
The operating system can see the status and is allowed to
request certain things, for example entering and exiting the
modes, but the support for alternate modes is very limited
in UCSI. The feature is also optional, which means that even
when the platform supports alternate modes, the operating
system may not be even made aware of them.

UCSI does not support direct VDM reading or writing.
Instead, alternate modes can be entered and exited using a
single custom command which takes also an optional SVID
specific configuration value as parameter. That means every
supported alternate mode has to be handled separately in
UCSI driver.

This commit does not include support for any specific
alternate mode. The discovered alternate modes are now
registered, but binding a driver to an alternate mode will
not be possible until support for that alternate mode is
added to the UCSI driver.

Tested-by: Ajay Gupta <ajayg@nvidia.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: ucsi: ccg: add firmware flashing support
Ajay Gupta [Tue, 23 Apr 2019 14:21:47 +0000 (17:21 +0300)]
usb: typec: ucsi: ccg: add firmware flashing support

CCGx has two copies of the firmware in addition to the bootloader.
If the device is running FW1, FW2 can be updated with the new version.
Dual firmware mode allows the CCG device to stay in a PD contract and
support USB PD and Type-C functionality while a firmware update is in
progress.

First we read the currently flashed firmware version of both
primary and secondary firmware and then compare it with
version of firmware file to determine if flashing is required.

Command framework is added to support sending commands to CCGx
controller. We wait for response after sending the command and then
read the response from RAB_RESPONSE register.

Below commands are supported,
- ENTER_FLASHING
- RESET
- PDPORT_ENABLE
- JUMP_TO_BOOT
- FLASH_ROW_RW
- VALIDATE_FW

Command specific mutex lock is also added to sync between driver
and user threads.

PD port number information is added which is required while sending
PD_PORT_ENABLE command

Signed-off-by: Ajay Gupta <ajayg@nvidia.com>
[ heikki: Added ABI documentation. ]
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoi2c: nvidia-gpu: Supply CCGx driver the fw build info
Ajay Gupta [Tue, 23 Apr 2019 14:21:46 +0000 (17:21 +0300)]
i2c: nvidia-gpu: Supply CCGx driver the fw build info

Adding device property "ccgx,firmware-build" for the CCGx
device, so the CCGx driver knows which firmware binary to
use for a specific vendor.

Suggested-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Ajay Gupta <ajayg@nvidia.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: ucsi: ccg: add get_fw_info function
Ajay Gupta [Tue, 23 Apr 2019 14:21:45 +0000 (17:21 +0300)]
usb: typec: ucsi: ccg: add get_fw_info function

Function is to get the details of ccg firmware and device version.
It will be useful in debugging and also during firmware update.

Signed-off-by: Ajay Gupta <ajayg@nvidia.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: usb251xb: Lock i2c-bus segment the hub resides
Serge Semin [Wed, 24 Apr 2019 14:49:14 +0000 (17:49 +0300)]
usb: usb251xb: Lock i2c-bus segment the hub resides

SMBus slave configuration is activated by CFG_SEL[1:0]=0x1 pins
state. This is the mode the hub is supposed to be to let this driver
work correctly. But a race condition might happen right after reset
is cleared due to CFG_SEL[0] pin being multiplexed with SMBus SCL
function. In case if the reset pin is handled by a i2c GPIO expander,
which is also placed at the same i2c-bus segment as the usb251x
SMB-interface connected to, then the hub reset clearance might
cause the CFG_SEL[0] being latched in unpredictable state. So
sometimes the hub configuration mode might be 0x1 (as expected),
but sometimes being 0x0, which doesn't imply to have the hub SMBus-slave
interface activated and consequently causes this driver failure.

In order to fix the problem we must make sure the GPIO-reset chip doesn't
reside the same i2c-bus segment as the SMBus-interface of the hub. If
it doesn't, we can safely block the segment for the time the reset is
cleared to prevent anyone generating a traffic at the i2c-bus SCL lane
connected to the CFG_SEL[0] pin. But if it does, nothing we can do, so
just return an error. If we locked the i2c-bus segment and tried to
communicate with the GPIO-expander, it would cause a deadlock. If we didn't
lock the i2c-bus segment, it would randomly cause the CFG_SEL[0] bit flip.

Signed-off-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: dwc3: Allow building USB_DWC3_QCOM without EXTCON
Marc Gonzalez [Wed, 24 Apr 2019 15:00:57 +0000 (17:00 +0200)]
usb: dwc3: Allow building USB_DWC3_QCOM without EXTCON

Keep EXTCON support optional, as some platforms do not need it.

Do the same for USB_DWC3_OMAP while we're at it.

Fixes: 3def4031b3e3f ("usb: dwc3: add EXTCON dependency for qcom")
Signed-off-by: Marc Gonzalez <marc.w.gonzalez@free.fr>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousbip: stub_rx: tidy the indenting in is_clear_halt_cmd()
Dan Carpenter [Wed, 24 Apr 2019 09:54:14 +0000 (12:54 +0300)]
usbip: stub_rx: tidy the indenting in is_clear_halt_cmd()

There is an extra space character before the return statement.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoMerge tag 'phy-for-5.2' of git://git.kernel.org/pub/scm/linux/kernel/git/kishon/linux...
Greg Kroah-Hartman [Thu, 25 Apr 2019 08:49:34 +0000 (10:49 +0200)]
Merge tag 'phy-for-5.2' of git://git./linux/kernel/git/kishon/linux-phy into usb-next

Kishon writes:

phy: for 5.2

  *) Add a new *release* phy_ops invoked when the consumer relinquishes PHY
     that can be used to undo the operation performed in xlate
  *) Add new driver to support USB2 PHY and shared USB3 + PCIE PHY in Amlogic
     G12A SoC Family.
  *) Add new driver to support for Broadcom's Stingray USB PHY (Type 1 has
     one super speed PHY and one high speed PHY, Type 2 has one high speed PHY)
  *) Add new driver to support USB PHY in hi3660 SoC of Hisilicon
  *) Add new driver to support UFS M-PHY in MediaTek SoC
  *) Add new driver to support XUSB pad controller in Tegra186 SoCs
  *) Add new driver to support SERDES in TI's AM654 platform
  *) Add support for generation 2 USB2 PHY and gneration 3 USB2 PHY in r8a77470
     to phy-rcar-gen2.c and phy-rcar-gen3-usb2.c respectively
  *) Add support for PCIe QMP PHY support in msm8998 to phy-qcom-qmp.c
  *) Add support for SERDES6G in phy-ocelot-serdes.c
  *) Add support to set drive impedance from device tree in phy-rockchip-emmc.c
  *) Add support to power up/down the VBUS voltage rail in phy-fsl-imx8mq-usb.c
  *) Add support to shut off regulators that power UFS during system suspend
  *) Re-design phy-rcar-gen3-usb2.c to create separate PHY instances for each
     channel which helps to enable/disable interrupts for each instance
     independently
  *) Fix PCIe power up sequence to follow the TRM in order to ensure the DPLL &
     PHY operates correctly over the entire temperature range.
  *) Use devm_clk_get_optional to get optional clocks instead of adding
     custom error checks

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
* tag 'phy-for-5.2' of git://git.kernel.org/pub/scm/linux/kernel/git/kishon/linux-phy: (51 commits)
  dt-bindings: phy-qcom-qmp: Tweak qcom,msm8998-qmp-ufs-phy
  dt-bindings: phy-qcom-qmp: Add qcom,msm8998-qmp-pcie-phy
  phy: Add usb phy support for hi3660 Soc of Hisilicon
  dt-bindings: phy: Add support for HiSilicon's hi3660 USB PHY
  scsi: phy: mediatek: fix typo in author's email address
  phy: ocelot-serdes: Add support for SERDES6G muxing
  phy: fsl-imx8mq-usb: add support for VBUS power control
  dt-bindings: phy-imx8mq-usb: add optional vbus supply regulator
  phy: qcom-qmp: Add msm8998 PCIe QMP PHY support
  phy: ti: am654-serdes: Support all clksel values
  phy: ti: Add a new SERDES driver for TI's AM654x SoC
  dt-bindings: phy: ti: Add dt binding documentation for SERDES in AM654x SoC
  phy: core: Invoke pm_runtime_get_*/pm_runtime_put_* before invoking reset callback
  phy: core: Add *release* phy_ops invoked when the consumer relinquishes PHY
  phy: phy-meson-gxl-usb2: get optional clock by devm_clk_get_optional()
  phy: socionext: get optional clock by devm_clk_get_optional()
  phy: qcom-qusb2: get optional clock by devm_clk_get_optional()
  phy: phy-mtk-tphy: get optional clock by devm_clk_get_optional()
  phy: renesas: rcar-gen3-usb2: enable/disable independent irqs
  phy: renesas: rcar-gen3-usb2: Use pdev's device pointer on dev_vdbg()
  ...

5 years agodt-bindings: phy-qcom-qmp: Tweak qcom,msm8998-qmp-ufs-phy
Marc Gonzalez [Mon, 1 Apr 2019 14:42:24 +0000 (16:42 +0200)]
dt-bindings: phy-qcom-qmp: Tweak qcom,msm8998-qmp-ufs-phy

Fixup MSM8998 UFS binding now that Evan's reset series has landed.
https://lore.kernel.org/lkml/20190321171800.104681-1-evgreen@chromium.org/

Signed-off-by: Marc Gonzalez <marc.w.gonzalez@free.fr>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agousb: mtu3: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 17 Apr 2019 08:28:21 +0000 (16:28 +0800)]
usb: mtu3: get optional clock by devm_clk_get_optional()

Use devm_clk_get_optional() to get optional clock instead of
optional_clk_get() which uses devm_clk_get() to get clock and
checks for -EPROBE_DEFER but not -ENOENT as devm_clk_get_optional()
does, in fact, only ignoring -ENOENT will cover more errors, so
the replacement doesn't change original purpose.

Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: chipidea: msm: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 17 Apr 2019 08:28:20 +0000 (16:28 +0800)]
usb: chipidea: msm: get optional clock by devm_clk_get_optional()

When the driver tries to get optional clock, it ignores all errors except
-EPROBE_DEFER, but if only ignores -ENOENT, it will cover some real errors,
such as -ENOMEM, so use devm_clk_get_optional() to get optional clock.

Cc: Peter Chen <Peter.Chen@nxp.com>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Acked-by: Peter Chen <Peter.Chen@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: dwc2: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 17 Apr 2019 08:28:19 +0000 (16:28 +0800)]
usb: dwc2: get optional clock by devm_clk_get_optional()

When the driver tries to get optional clock, it ignores all errors,
but if only ignores -ENOENT, it will cover some real errors, such as
-EPROBE_DEFER, so use devm_clk_get_optional() to get optional clock.

Cc: Minas Harutyunyan <hminas@synopsys.com>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Acked-by: Minas Harutyunyan <hminas@synopsys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: misc: usb3503: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 17 Apr 2019 08:28:18 +0000 (16:28 +0800)]
usb: misc: usb3503: get optional clock by devm_clk_get_optional()

When the driver tries to get optional clock, it ignores all errors except
-EPROBE_DEFER, but if only ignores -ENOENT, it will cover some real errors,
such as -ENOMEM, so use devm_clk_get_optional() to get optional clock.
And remove unnecessary stack variable clk.

Cc: Dongjin Kim <tobetter@gmail.com>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: host: xhci-plat: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 17 Apr 2019 08:28:17 +0000 (16:28 +0800)]
usb: host: xhci-plat: get optional clock by devm_clk_get_optional()

When the driver tries to get optional clock, it ignores all errors except
-EPROBE_DEFER, but if only ignores -ENOENT, it will cover some real errors,
such as -ENOMEM, so use devm_clk_get_optional() to get optional clock.

Cc: Mathias Nyman <mathias.nyman@intel.com>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: xhci-mtk: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 17 Apr 2019 08:28:16 +0000 (16:28 +0800)]
usb: xhci-mtk: get optional clock by devm_clk_get_optional()

Use devm_clk_get_optional() to get optional clock instead of
optional_clk_get() which uses devm_clk_get() to get clock and
checks for -EPROBE_DEFER but not -ENOENT as devm_clk_get_optional()
does, in fact, only ignoring -ENOENT will cover more errors, so the
replacement doesn't change original purpose.

Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: fusb302: Revert "Resolve fixed power role contract setup"
Hans de Goede [Tue, 16 Apr 2019 20:07:54 +0000 (22:07 +0200)]
usb: typec: fusb302: Revert "Resolve fixed power role contract setup"

Some tcpc device-drivers need to explicitly be told to watch for connection
events, otherwise the tcpc will not generate any TCPM_CC_EVENTs and devices
being plugged into the Type-C port will not be noticed.

For dual-role ports tcpm_start_drp_toggling() is used to tell the tcpc to
watch for connection events. But for single-role ports we've so far been
falling back to just calling tcpm_set_cc(). For some tcpc-s such as the
fusb302 this is not enough and no TCPM_CC_EVENT will be generated.

Commit ea3b4d5523bc ("usb: typec: fusb302: Resolve fixed power role
contract setup") fixed SRPs not working because of this by making the
fusb302 driver start connection detection on every tcpm_set_cc() call.
It turns out this breaks src->snk power-role swapping because during the
swap we first set the Cc pins to Rp, calling set_cc, and then send a PS_RDY
message. But the fusb302 cannot send PD messages while its toggling engine
is active, so sending the PS_RDY message fails.

Struct tcpc_dev now has a new start_srp_connection_detect callback and
fusb302.c now implements this. This callback gets called when we the
fusb302 needs to start connection detection, fixing fusb302 SRPs not
seeing connected devices.

This allows us to revert the changes to fusb302's set_cc implementation,
making it once again purely setup the Cc-s and matching disconnect
detection, fixing src->snk power-role swapping no longer working.

Note that since the code was refactored in between, codewise this is not a
straight forward revert. Functionality wise this is a straight revert and
the original functionality is fully restored.

Fixes: ea3b4d5523bc ("usb: typec: fusb302: Resolve fixed power role ...")
Cc: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Tested-by: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: fusb302: Implement start_toggling for all port-types
Hans de Goede [Tue, 16 Apr 2019 20:07:53 +0000 (22:07 +0200)]
usb: typec: fusb302: Implement start_toggling for all port-types

When in single-role port mode, we must start single-role toggling to
get an interrupt when a device / cable gets plugged into the port.

This commit modifies the fusb302 start_toggling implementation to
start toggling for all port-types, so that connection-detection works
on single-role ports too.

Fixes: ea3b4d5523bc("usb: typec: fusb302: Resolve fixed power role ...")
Cc: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Tested-by: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: typec: tcpm: Notify the tcpc to start connection-detection for SRPs
Hans de Goede [Tue, 16 Apr 2019 20:07:52 +0000 (22:07 +0200)]
usb: typec: tcpm: Notify the tcpc to start connection-detection for SRPs

Some tcpc device-drivers need to explicitly be told to watch for connection
events, otherwise the tcpc will not generate any TCPM_CC_EVENTs and devices
being plugged into the Type-C port will not be noticed.

For dual-role ports tcpm_start_drp_toggling() is used to tell the tcpc to
watch for connection events. Sofar we lack a similar callback to the tcpc
for single-role ports. With some tcpc-s such as the fusb302 this means
no TCPM_CC_EVENTs will be generated when the port is configured as a
single-role port.

This commit renames start_drp_toggling to start_toggling and since the
device-properties are parsed by the tcpm-core, adds a port_type parameter
to the start_toggling callback so that the tcpc_dev driver knows the
port-type and can act accordingly when it starts toggling.

The new start_toggling callback now always gets called if defined, instead
of only being called for DRP ports.

To avoid this causing undesirable functional changes all existing
start_drp_toggling implementations are not only renamed to start_toggling,
but also get a port_type check added and return -EOPNOTSUPP when port_type
is not DRP.

Fixes: ea3b4d5523bc("usb: typec: fusb302: Resolve fixed power role ...")
Cc: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Tested-by: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: host: use usb_endpoint_maxp instead of usb_maxpacket
Yan Zhu [Thu, 18 Apr 2019 15:25:21 +0000 (23:25 +0800)]
usb: host: use usb_endpoint_maxp instead of usb_maxpacket

fhci_queue_urb() shouldn't use urb->pipe to compute the maxpacket
size anyway.It should use usb_endpoint_maxp(&urb->ep->desc).

Signed-off-by: Yan Zhu <zhuyan34@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: phy-qcom-qmp: Add qcom,msm8998-qmp-pcie-phy
Marc Gonzalez [Tue, 9 Apr 2019 12:49:00 +0000 (14:49 +0200)]
dt-bindings: phy-qcom-qmp: Add qcom,msm8998-qmp-pcie-phy

Add compatible string for QMP PCIe phy on msm8998.

Signed-off-by: Marc Gonzalez <marc.w.gonzalez@free.fr>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: Add usb phy support for hi3660 Soc of Hisilicon
Yu Chen [Fri, 29 Mar 2019 04:14:03 +0000 (12:14 +0800)]
phy: Add usb phy support for hi3660 Soc of Hisilicon

This driver handles usb phy power on and shutdown for hi3660 Soc of
Hisilicon.

Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Kishon Vijay Abraham I <kishon@ti.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Shawn Guo <shawnguo@kernel.org>
Cc: Pengcheng Li <lpc.li@hisilicon.com>
Cc: Jianguo Sun <sunjianguo1@huawei.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Jiancheng Xue <xuejiancheng@hisilicon.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Binghui Wang <wangbinghui@hisilicon.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Yu Chen <chenyu56@huawei.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: Add support for HiSilicon's hi3660 USB PHY
Yu Chen [Fri, 29 Mar 2019 04:13:57 +0000 (12:13 +0800)]
dt-bindings: phy: Add support for HiSilicon's hi3660 USB PHY

This patch adds binding documentation for supporting the hi3660 usb
phy on boards like the HiKey960.

Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Binghui Wang <wangbinghui@hisilicon.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Yu Chen <chenyu56@huawei.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agoscsi: phy: mediatek: fix typo in author's email address
Colin Ian King [Wed, 20 Mar 2019 21:22:04 +0000 (21:22 +0000)]
scsi: phy: mediatek: fix typo in author's email address

There is a typo in the module author's email address. Fix this.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Acked-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ocelot-serdes: Add support for SERDES6G muxing
Kavya Sree Kotagiri [Mon, 25 Mar 2019 10:13:33 +0000 (10:13 +0000)]
phy: ocelot-serdes: Add support for SERDES6G muxing

Adding support for SERDES6G muxing required for QSGMII mode of operation.

Signed-off-by: Kavya Sree Kotagiri <kavyasree.kotagiri@microchip.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: Steen Hegelund <Steen.Hegelund@microchip.com>
Co-developed-by: Quentin Schulz <quentin.schulz@bootlin.com>
Co-developed-by: Steen Hegelund <Steen.Hegelund@microchip.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: fsl-imx8mq-usb: add support for VBUS power control
Lucas Stach [Thu, 4 Apr 2019 16:41:48 +0000 (18:41 +0200)]
phy: fsl-imx8mq-usb: add support for VBUS power control

This adds support to the PHY driver to power up/down the VBUS
voltage rail at the appropriate times.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy-imx8mq-usb: add optional vbus supply regulator
Lucas Stach [Thu, 4 Apr 2019 16:41:47 +0000 (18:41 +0200)]
dt-bindings: phy-imx8mq-usb: add optional vbus supply regulator

Add a vbus supply regulator phandle, so the PHY can enable the VBUS
voltage rail when powering up.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agoUSB: core: Don't unbind interfaces following device reset failure
Alan Stern [Tue, 16 Apr 2019 14:50:01 +0000 (10:50 -0400)]
USB: core: Don't unbind interfaces following device reset failure

The SCSI core does not like to have devices or hosts unregistered
while error recovery is in progress.  Trying to do so can lead to
self-deadlock: Part of the removal code tries to obtain a lock already
held by the error handler.

This can cause problems for the usb-storage and uas drivers, because
their error handler routines perform a USB reset, and if the reset
fails then the USB core automatically goes on to unbind all drivers
from the device's interfaces -- all while still in the context of the
SCSI error handler.

As it turns out, practically all the scenarios leading to a USB reset
failure end up causing a device disconnect (the main error pathway in
usb_reset_and_verify_device(), at the end of the routine, calls
hub_port_logical_disconnect() before returning).  As a result, the
hub_wq thread will soon become aware of the problem and will unbind
all the device's drivers in its own context, not in the
error-handler's context.

This means that usb_reset_device() does not need to call
usb_unbind_and_rebind_marked_interfaces() in cases where
usb_reset_and_verify_device() has returned an error, because hub_wq
will take care of everything anyway.

This particular problem was observed in somewhat artificial
circumstances, by using usbfs to tell a hub to power-down a port
connected to a USB-3 mass storage device using the UAS protocol.  With
the port turned off, the currently executing command timed out and the
error handler started running.  The USB reset naturally failed,
because the hub port was off, and the error handler deadlocked as
described above.  Not carrying out the call to
usb_unbind_and_rebind_marked_interfaces() fixes this issue.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: Kento Kobayashi <Kento.A.Kobayashi@sony.com>
Tested-by: Kento Kobayashi <Kento.A.Kobayashi@sony.com>
CC: Bart Van Assche <bvanassche@acm.org>
CC: Martin K. Petersen <martin.petersen@oracle.com>
CC: Jacky Cao <Jacky.Cao@sony.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agophy: qcom-qmp: Add msm8998 PCIe QMP PHY support
Marc Gonzalez [Tue, 9 Apr 2019 12:48:22 +0000 (14:48 +0200)]
phy: qcom-qmp: Add msm8998 PCIe QMP PHY support

Documentation for this PHY, and the proper configuration settings,
is *not* publicly available. Therefore the initialization sequence
is copied wholesale from downstream:

https://source.codeaurora.org/quic/la/kernel/msm-4.4/tree/arch/arm/boot/dts/qcom/msm8998-v2.dtsi?h=LE.UM.1.3.r3.25#n372

Reviewed-by: Vivek Gautam <vivek.gautam@codeaurora.org>
Signed-off-by: Marc Gonzalez <marc.w.gonzalez@free.fr>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti: am654-serdes: Support all clksel values
Roger Quadros [Fri, 5 Apr 2019 11:08:34 +0000 (16:38 +0530)]
phy: ti: am654-serdes: Support all clksel values

Add support to select all 16 CLKSEL combinations that are shown in
"SerDes Reference Clock Distribution" in AM65 TRM.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti: Add a new SERDES driver for TI's AM654x SoC
Kishon Vijay Abraham I [Wed, 17 Apr 2019 06:19:39 +0000 (11:49 +0530)]
phy: ti: Add a new SERDES driver for TI's AM654x SoC

Add a new SERDES driver for TI's AM654x SoC which configures
the SERDES only for PCIe. Support fo USB3 will be added later.

SERDES in am654x has three input clocks (left input, externel reference
clock and right input) and two output clocks (left output and right
output) in addition to a PLL mux clock which the SERDES uses for Clock
Multiplier Unit (CMU refclock).

The PLL mux clock can select from one of the three input clocks.
The right output can select between left input and external reference
clock while the left output can select between the right input and
external reference clock.

The driver has support to select PLL mux and left/right output mux as
specified in device tree.

[rogerq@ti.com: Fix boot lockup caused by accessing a structure member
(hw->init) allocated in stack of probe() and accessed in get_parent]
[rogerq@ti.com: Fix "Failed to find the parent" warnings]
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: ti: Add dt binding documentation for SERDES in AM654x SoC
Kishon Vijay Abraham I [Fri, 5 Apr 2019 11:08:32 +0000 (16:38 +0530)]
dt-bindings: phy: ti: Add dt binding documentation for SERDES in AM654x SoC

AM654x has two SERDES instances. Each instance has three input clocks
(left input, externel reference clock and right input) and two output
clocks (left output and right output) in addition to a PLL mux clock
which the SERDES uses for Clock Multiplier Unit (CMU refclock).
The PLL mux clock can select from one of the three input clocks.
The right output can select between left input and external reference
clock while the left output can select between the right input and
external reference clock.

The left and right input reference clock of SERDES0 and SERDES1
respectively are connected to the SoC clock. In the case of two lane
SERDES personality card, the left input of SERDES1 is connected to
the right output of SERDES0 in a chained fashion.

See section "Reference Clock Distribution" of AM65x Sitara Processors
TRM (SPRUID7 – April 2018) for more details.

Add dt-binding documentation in order to represent all these different
configurations in device tree.

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: core: Invoke pm_runtime_get_*/pm_runtime_put_* before invoking reset callback
Kishon Vijay Abraham I [Fri, 5 Apr 2019 11:08:31 +0000 (16:38 +0530)]
phy: core: Invoke pm_runtime_get_*/pm_runtime_put_* before invoking reset callback

PHY drivers may try to access PHY registers in the ->reset() callback.
Invoke phy_pm_runtime_get_sync() before invoking the ->reset() callback
so that the PHY drivers don't have to enable clocks by themselves before
accessing PHY registers.

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: core: Add *release* phy_ops invoked when the consumer relinquishes PHY
Kishon Vijay Abraham I [Fri, 5 Apr 2019 11:08:30 +0000 (16:38 +0530)]
phy: core: Add *release* phy_ops invoked when the consumer relinquishes PHY

Add a new phy_ops *release* invoked when the consumer relinquishes the
PHY using phy_put/devm_phy_put. The initializations done by the PHY
driver in of_xlate call back can be can be cleaned up here.

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: phy-meson-gxl-usb2: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 10 Apr 2019 06:13:06 +0000 (14:13 +0800)]
phy: phy-meson-gxl-usb2: get optional clock by devm_clk_get_optional()

Use devm_clk_get_optional() to get optional clock

Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Acked-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: socionext: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 10 Apr 2019 06:13:05 +0000 (14:13 +0800)]
phy: socionext: get optional clock by devm_clk_get_optional()

Use devm_clk_get_optional() to get optional clock

Cc: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Reviewed-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: qcom-qusb2: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 10 Apr 2019 06:13:04 +0000 (14:13 +0800)]
phy: qcom-qusb2: get optional clock by devm_clk_get_optional()

Use devm_clk_get_optional() to get optional clock

Cc: Andy Gross <andy.gross@linaro.org>
Cc: David Brown <david.brown@linaro.org>
Cc: Vivek Gautam <vivek.gautam@codeaurora.org>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: phy-mtk-tphy: get optional clock by devm_clk_get_optional()
Chunfeng Yun [Wed, 10 Apr 2019 06:13:03 +0000 (14:13 +0800)]
phy: phy-mtk-tphy: get optional clock by devm_clk_get_optional()

Use devm_clk_get_optional() to get optional clock

Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: renesas: rcar-gen3-usb2: enable/disable independent irqs
Yoshihiro Shimoda [Thu, 11 Apr 2019 10:27:36 +0000 (19:27 +0900)]
phy: renesas: rcar-gen3-usb2: enable/disable independent irqs

Since the previous code enabled/disabled the irqs both OHCI and EHCI,
it is possible to cause unexpected interruptions. To avoid this,
this patch creates multiple phy instances from phandle and
enables/disables independent irqs by the instances.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com>
Tested-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: renesas: rcar-gen3-usb2: Use pdev's device pointer on dev_vdbg()
Yoshihiro Shimoda [Thu, 11 Apr 2019 10:27:35 +0000 (19:27 +0900)]
phy: renesas: rcar-gen3-usb2: Use pdev's device pointer on dev_vdbg()

To implement multiple phy instances in the future, this patch uses
pdev's device pointer on dev_vdbg() instead of the phy's device
pointer.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: rcar-gen3-phy-usb2: Revise #phy-cells property
Yoshihiro Shimoda [Thu, 11 Apr 2019 10:27:34 +0000 (19:27 +0900)]
dt-bindings: phy: rcar-gen3-phy-usb2: Revise #phy-cells property

To have the detailed property on each PHY specifier, this patch revises
the #phy-cells property.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: rcar-gen3-usb2: Add support for r8a77470
Biju Das [Wed, 10 Apr 2019 14:48:41 +0000 (15:48 +0100)]
phy: rcar-gen3-usb2: Add support for r8a77470

This patch adds support for r8a77470 (RZ/G1C). We can reuse this driver for
initializing timing/interrupt generation registers.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: renesas: phy-rcar-gen2: Add support for r8a77470
Biju Das [Wed, 10 Apr 2019 14:48:39 +0000 (15:48 +0100)]
phy: renesas: phy-rcar-gen2: Add support for r8a77470

This patch adds support for RZ/G1C (r8a77470) SoC. RZ/G1C SoC has a
PLL register shared between hsusb0 and hsusb1. Compared to other RZ/G1
and R-Car Gen2/3, USB Host needs to deassert the pll reset.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-and-Tested-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: rcar-gen3-phy-usb2: Add r8a77470 support
Biju Das [Wed, 10 Apr 2019 14:48:40 +0000 (15:48 +0100)]
dt-bindings: rcar-gen3-phy-usb2: Add r8a77470 support

Document RZ/G1C (R8A77470) SoC bindings.

For RZ/G1C, this driver is used to enable interrupt generation and
initializing timing registers which is part of phy_init code.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: rcar-gen2: Add r8a77470 support
Biju Das [Wed, 10 Apr 2019 14:48:38 +0000 (15:48 +0100)]
dt-bindings: phy: rcar-gen2: Add r8a77470 support

Add USB PHY support for r8a77470 SoC. Renesas RZ/G1C (R8A77470)
USB PHY is similar to the R-Car Gen2 family, but has the below
feature compared to other RZ/G1 and R-Car Gen2/3 SoCs

It has a shared pll reset for usbphy0/usbphy1 and this register
reside in usbphy0 block.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti-pipe3: Fix PCIe power up sequence
Roger Quadros [Fri, 22 Mar 2019 08:58:07 +0000 (10:58 +0200)]
phy: ti-pipe3: Fix PCIe power up sequence

TRM [1] mentions that we need to power up
PCIESS_PHY_TX and PCIESS_PHY_RX before configuring
PCIe_PHY_RX SCP settings.

See "Table 26-81. PCIePHY Subsystem Low-Level Programming Sequence".

[1] DRA75x, DRA74x TRM - http://www.ti.com/lit/ug/sprui30f/sprui30f.pdf

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti-pipe3: Fix SATA & USB PHY power up sequence
Roger Quadros [Fri, 22 Mar 2019 08:58:06 +0000 (10:58 +0200)]
phy: ti-pipe3: Fix SATA & USB PHY power up sequence

As per "Table 26-7. SATA PHY Subsystem Low-Level Programming Sequence"
in TRM [1] we need to turn on SATA_PHY_TX before SATA_PHY_RX.

[1] DRA75x, DRA74x TRM - http://www.ti.com/lit/ug/sprui30f/sprui30f.pdf

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti-pipe3: improve DPLL stability for SATA & USB
Roger Quadros [Fri, 22 Mar 2019 08:58:05 +0000 (10:58 +0200)]
phy: ti-pipe3: improve DPLL stability for SATA & USB

For increased DPLL stability use the settings recommended in
the TRM [1] for PHY_RX registers for SATA and USB.

For SATA we need to use spread spectrum settings even
though we don't have spread spectrum enabled. The
suggested non-spread spectrum settings don't work.

[1] DRA75x, DRA74x TRM - http://www.ti.com/lit/ug/sprui30f/sprui30f.pdf

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti-pipe3: Introduce mode property in driver data
Roger Quadros [Fri, 22 Mar 2019 08:58:04 +0000 (10:58 +0200)]
phy: ti-pipe3: Introduce mode property in driver data

Introduce a mode property in the driver data so that
we don't have to keep using "of_device_is_compatible()"
throughtout the driver.

No functional change.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ti-pipe3: fix missing bit-wise or operator when assigning val
Colin Ian King [Tue, 19 Feb 2019 14:53:49 +0000 (14:53 +0000)]
phy: ti-pipe3: fix missing bit-wise or operator when assigning val

There seems to be a missing bit-wise or operator when setting val,
fix this by adding it in.

Fixes: 2796ceb0c18a ("phy: ti-pipe3: Update pcie phy settings")
Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: mediatek: Add UFS M-PHY driver
Stanley Chu [Sat, 16 Mar 2019 05:04:46 +0000 (13:04 +0800)]
phy: mediatek: Add UFS M-PHY driver

Add UFS M-PHY driver on MediaTek chipsets.

Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: Add document for phy-mtk-ufs
Stanley Chu [Sat, 16 Mar 2019 05:04:44 +0000 (13:04 +0800)]
dt-bindings: phy: Add document for phy-mtk-ufs

Add UFS M-PHY node document for MediaTek SoC chips.

Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: Add a new property drive-impedance-ohm for RK's emmc PHY
Christoph Muellner [Fri, 22 Mar 2019 11:34:51 +0000 (12:34 +0100)]
dt-bindings: phy: Add a new property drive-impedance-ohm for RK's emmc PHY

This patch documents the new proprty drive-impedance-ohm for
Rockchip's eMMC PHY node.

Signed-off-by: Christoph Muellner <christoph.muellner@theobroma-systems.com>
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: rockchip-emmc: Allow to set drive impedance via DTS.
Christoph Muellner [Fri, 22 Mar 2019 11:34:50 +0000 (12:34 +0100)]
phy: rockchip-emmc: Allow to set drive impedance via DTS.

The rockchip-emmc PHY can be configured with different
drive impedance values. Currenlty a value of 50 Ohm is
hard coded into the driver.

This patch introduces the DTS property 'drive-impedance-ohm'
for the rockchip-emmc phy node, which uses the value from the DTS
to setup the drive impedance accordingly.

Signed-off-by: Christoph Muellner <christoph.muellner@theobroma-systems.com>
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: ufs-qcom: Refactor all init steps into phy_poweron
Evan Green [Thu, 21 Mar 2019 17:18:00 +0000 (10:18 -0700)]
phy: ufs-qcom: Refactor all init steps into phy_poweron

The phy code was using implicit sequencing between the PHY driver
and the UFS driver to implement certain hardware requirements.
Specifically, the PHY reset register in the UFS controller needs
to be deasserted before serdes start occurs in the PHY.

Before this change, the code was doing this by utilizing the two
phy callbacks, phy_init() and phy_poweron(), as "init step 1" and
"init step 2", where the UFS driver would deassert reset between
these two steps.

This makes it challenging to power off the regulators in suspend,
as regulators are initialized in init, not in poweron(), but only
poweroff() is called during suspend, not exit().

For UFS, move the actual firing up of the PHY to phy_poweron() and
phy_poweroff() callbacks, rather than init()/exit(). UFS calls
phy_poweroff() during suspend, so now all clocks and regulators for
the phy can be powered down during suspend.

QMP is a little tricky because the PHY is also shared with PCIe and
USB3, which have their own definitions for init() and poweron(). Rename
the meaty functions to _enable() and _disable() to disentangle from the
PHY core names, and then create two different ops structures: one for
UFS and one for the other PHY types.

In phy-qcom-ufs, remove the 'is_powered_on' and 'is_started' guards,
as the generic PHY code does the reference counting. The
14/20nm-specific init functions get collapsed into the generic power_on()
function, with the addition of a calibrate() callback specific to 14/20nm.

Signed-off-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: qcom: Utilize UFS reset controller
Evan Green [Thu, 21 Mar 2019 17:17:59 +0000 (10:17 -0700)]
phy: qcom: Utilize UFS reset controller

Move the PHY reset from ufs-qcom into the respective PHYs. This will
allow us to merge the two phases of UFS PHY initialization.

Signed-off-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agoscsi: ufs: qcom: Expose the reset controller for PHY
Evan Green [Thu, 21 Mar 2019 17:17:58 +0000 (10:17 -0700)]
scsi: ufs: qcom: Expose the reset controller for PHY

Expose a reset controller that the phy will later use to control its
own PHY reset in the UFS controller. This will enable the combining
of PHY init functionality into a single function.

Signed-off-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: qcom-ufs: Add resets property
Evan Green [Thu, 21 Mar 2019 17:17:55 +0000 (10:17 -0700)]
dt-bindings: phy: qcom-ufs: Add resets property

Add a resets property to the PHY that represents the PHY reset
register in the UFS controller itself. This better describes the
complete specification of the PHY, and allows the PHY to perform
its initialization in a single function, rather than relying on
back-channel sequencing of initialization through the PHY framework.

Signed-off-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy-qcom-qmp: Add UFS PHY reset
Evan Green [Thu, 21 Mar 2019 17:17:54 +0000 (10:17 -0700)]
dt-bindings: phy-qcom-qmp: Add UFS PHY reset

Add a required reset to the SDM845 UFS phy to express the PHY reset
bit inside the UFS controller register space. Before this change, this
reset was not expressed in the DT, and the driver utilized two different
callbacks (phy_init and phy_poweron) to implement a two-phase
initialization procedure that involved deasserting this reset between
init and poweron. This abused the two callbacks and diluted their
purpose.

That scheme does not work as regulators cannot be turned off in
phy_poweroff because they were turned on in init, rather than poweron.
The net result is that regulators are left on in suspend that shouldn't
be.

This new scheme gives the UFS reset to the PHY, so that it can fully
initialize itself in a single callback. We can then turn regulators on
during poweron and off during poweroff.

Signed-off-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: ufs: Add #reset-cells for Qualcomm controllers
Evan Green [Thu, 21 Mar 2019 17:17:53 +0000 (10:17 -0700)]
dt-bindings: ufs: Add #reset-cells for Qualcomm controllers

Enable Qualcomm UFS controllers to expose the PHY reset via a reset
controller.

Signed-off-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: amlogic: Add Amlogic G12A USB3 + PCIE Combo PHY Driver
Neil Armstrong [Mon, 25 Mar 2019 09:39:41 +0000 (10:39 +0100)]
phy: amlogic: Add Amlogic G12A USB3 + PCIE Combo PHY Driver

This adds support for the shared USB3 + PCIE PHY found in the
Amlogic G12A SoC Family.

It supports USB3 Host mode or PCIE 2.0 mode, depending on the layout of
the board.

Selection is done by the #phy-cells, making the mode static and exclusive.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: amlogic: add Amlogic G12A USB2 PHY Driver
Neil Armstrong [Mon, 25 Mar 2019 09:39:40 +0000 (10:39 +0100)]
phy: amlogic: add Amlogic G12A USB2 PHY Driver

This adds support for the USB2 PHY found in the Amlogic G12A SoC Family.

It supports Host and/or Peripheral mode, depending on it's position.
The first PHY is only used as Host, but the second supports Dual modes
defined by the USB Control Glue HW in front of the USB Controllers.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: Add Amlogic G12A USB3+PCIE Combo PHY Bindings
Neil Armstrong [Mon, 25 Mar 2019 09:39:37 +0000 (10:39 +0100)]
dt-bindings: phy: Add Amlogic G12A USB3+PCIE Combo PHY Bindings

Add the Amlogic G12A Family USB3 + PCIE Combo PHY Bindings.

This PHY can provide exclusively USB3 or PCIE support on shared I/Os.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: Add Amlogic G12A USB2 PHY Bindings
Neil Armstrong [Mon, 25 Mar 2019 09:39:36 +0000 (10:39 +0100)]
dt-bindings: phy: Add Amlogic G12A USB2 PHY Bindings

Add the Amlogic G12A Family USB2 OTG PHY Bindings

The PHY can work in host or peripheral modes depending on it's position.
Configuration of the mode is part of the USBCTRL registers which are
outside of the PHY registers.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: sr-usb: Add Stingray USB PHY driver
Srinath Mannam [Tue, 19 Mar 2019 09:15:43 +0000 (14:45 +0530)]
phy: sr-usb: Add Stingray USB PHY driver

USB PHY driver supports two types of stingray USB PHYs
 - Type 1 is a combo PHY contains two PHYs, one SS and one HS.
 - Type 2 is a single HS PHY.

These two PHY versons support both Generic xHCI host controller driver
and BDC Broadcom device controller driver.

Signed-off-by: Srinath Mannam <srinath.mannam@broadcom.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: Add Stingray USB PHY binding document
Srinath Mannam [Tue, 19 Mar 2019 09:15:42 +0000 (14:45 +0530)]
dt-bindings: phy: Add Stingray USB PHY binding document

Add DT binding document for Stingray USB PHY.

Signed-off-by: Srinath Mannam <srinath.mannam@broadcom.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: tegra: xusb: Add Tegra186 support
JC Kuo [Thu, 21 Feb 2019 15:46:34 +0000 (16:46 +0100)]
phy: tegra: xusb: Add Tegra186 support

Add support for the XUSB pad controller found on Tegra186 SoCs. It is
mostly similar to the same IP found on earlier chips, but the number of
pads exposed differs, as do the programming sequences.

Note that the DVDD_PEX, DVDD_PEX_PLL, HVDD_PEX and HVDD_PEX_PLL power
supplies of the XUSB pad controller require strict power sequencing and
are therefore controlled by the PMIC on Tegra186.

Signed-off-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
[dan.carpenter@oracle.com: Fix testing the wrong variable in probe()]
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
[yuehaibing@huawei.com: Make two functions static to fix sparse warning]
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: tegra: xusb: Add support for power supplies
Thierry Reding [Thu, 21 Feb 2019 15:46:33 +0000 (16:46 +0100)]
phy: tegra: xusb: Add support for power supplies

Support enabling various supplies needed to provide power to the PLLs
and logic used to drive the USB, PCI and SATA pads.

Reviewed-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: tegra: xusb: Parse dual-role mode property
Thierry Reding [Thu, 21 Feb 2019 15:46:32 +0000 (16:46 +0100)]
phy: tegra: xusb: Parse dual-role mode property

The device tree bindings document the "mode" property of "ports"
subnodes, but the driver was not parsing the property. In preparation
for adding role switching, parse the property at probe time.

Based on work by JC Kuo <jckuo@nvidia.com>.

Reviewed-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agophy: tegra: xusb: Skip single function lane programming
JC Kuo [Thu, 21 Feb 2019 15:46:31 +0000 (16:46 +0100)]
phy: tegra: xusb: Skip single function lane programming

Tegra186 USB2 pads and USB3 pads do not have hardware mux for changing
the pad function. For such "lanes", we can skip the lane mux register
programming.

Signed-off-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agodt-bindings: phy: tegra: Add Tegra186 support
Thierry Reding [Thu, 21 Feb 2019 15:46:30 +0000 (16:46 +0100)]
dt-bindings: phy: tegra: Add Tegra186 support

Extend the bindings to cover the set of features found in Tegra186. Note
that, technically, there are four more supplies connected to the XUSB
pad controller (DVDD_PEX, DVDD_PEX_PLL, HVDD_PEX and HVDD_PEX_PLL), but
the power sequencing requirements of Tegra186 require these to be under
the control of the PMIC.

Reviewed-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agousb-storage: Set virt_boundary_mask to avoid SG overflows
Alan Stern [Mon, 15 Apr 2019 17:19:25 +0000 (13:19 -0400)]
usb-storage: Set virt_boundary_mask to avoid SG overflows

The USB subsystem has always had an unusual requirement for its
scatter-gather transfers: Each element in the scatterlist (except the
last one) must have a length divisible by the bulk maxpacket size.
This is a particular issue for USB mass storage, which uses SG lists
created by the block layer rather than setting up its own.

So far we have scraped by okay because most devices have a logical
block size of 512 bytes or larger, and the bulk maxpacket sizes for
USB 2 and below are all <= 512.  However, USB 3 has a bulk maxpacket
size of 1024.  Since the xhci-hcd driver includes native SG support,
this hasn't mattered much.  But now people are trying to use USB-3
mass storage devices with USBIP, and the vhci-hcd driver currently
does not have full SG support.

The result is an overflow error, when the driver attempts to implement
an SG transfer of 63 512-byte blocks as a single
3584-byte (7 blocks) transfer followed by seven 4096-byte (8 blocks)
transfers.  The device instead sends 31 1024-byte packets followed by
a 512-byte packet, and this overruns the first SG buffer.

Ideally this would be fixed by adding better SG support to vhci-hcd.
But for now it appears we can work around the problem by
asking the block layer to respect the maxpacket limitation, through
the use of the virt_boundary_mask.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: Seth Bollinger <Seth.Bollinger@digi.com>
Tested-by: Seth Bollinger <Seth.Bollinger@digi.com>
CC: Ming Lei <tom.leiming@gmail.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb: Convert the generic EHCI binding to YAML
Maxime Ripard [Tue, 16 Apr 2019 08:27:59 +0000 (10:27 +0200)]
dt-bindings: usb: Convert the generic EHCI binding to YAML

The generic EHCI binding is used by many controllers that are using the
EHCI spec.

Convert that binding to a YAML description to enable the validation on all
the nodes using that binding.

Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Acked-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb: Convert the generic OHCI binding to YAML
Maxime Ripard [Tue, 16 Apr 2019 08:27:58 +0000 (10:27 +0200)]
dt-bindings: usb: Convert the generic OHCI binding to YAML

The generic OHCI binding is used by many controllers that are using the
OHCI spec.

Convert that binding to a YAML description to enable the validation on all
the nodes using that binding.

Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Acked-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb: Convert USB HCD generic binding to YAML
Maxime Ripard [Tue, 16 Apr 2019 08:27:57 +0000 (10:27 +0200)]
dt-bindings: usb: Convert USB HCD generic binding to YAML

The USB HCD generic binding is used by many USB host bindings.

In order to allow the DT validation to happen on those, let's create a YAML
description for that generic binding that can be referenced later on.

Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Acked-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoUSB: hub: Remove returned value 'status' since never used
Mathieu Malaterre [Wed, 27 Mar 2019 19:40:34 +0000 (20:40 +0100)]
USB: hub: Remove returned value 'status' since never used

The returned value in status has never been used since
commit 4296c70a5ec3 ("USB/xHCI: Enable USB 3.0 hub remote wakeup.")
So remove 'status' completely.

Remove warning (W=1):

  drivers/usb/core/hub.c:3671:8: warning: variable 'status' set but not used [-Wunused-but-set-variable]

Signed-off-by: Mathieu Malaterre <malat@debian.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodocs: usb: convert documents to ReST
Mauro Carvalho Chehab [Tue, 16 Apr 2019 02:56:01 +0000 (23:56 -0300)]
docs: usb: convert documents to ReST

Convert USB documents to ReST, in order to prepare for adding it
to the kernel API book, as most of the stuff there are driver or
subsystem-related.

Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: host: xhci-tegra: Add Tegra186 XUSB support
JC Kuo [Mon, 1 Apr 2019 10:40:47 +0000 (12:40 +0200)]
usb: host: xhci-tegra: Add Tegra186 XUSB support

This commit adds Tegra186 XUSB host mode controller support. This is
very similar to the existing support for Tegra124 and Tegra210, except
that the number of ports and PHYs differs and the IPFS wrapper being
gone.

Signed-off-by: JC Kuo <jckuo@nvidia.com>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agousb: host: xhci-tegra: Selectively program IPFS
JC Kuo [Mon, 1 Apr 2019 10:40:46 +0000 (12:40 +0200)]
usb: host: xhci-tegra: Selectively program IPFS

Starting with Tegra186, the XUSB controller no longer has the IPFS
wrapper. This commit adds a "has_ipfs" field to struct tegra_xusb_soc
that can be used to declare the existence of the IPFS wrapper.

For the existing chips (i.e. Tegra124 and Tegra210), the new field is
set to true. A future patch adding support for Tegra186 will set it to
false.

Signed-off-by: JC Kuo <jckuo@nvidia.com>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agodt-bindings: usb: xhci-tegra: Add Tegra186 support
Thierry Reding [Mon, 1 Apr 2019 10:40:45 +0000 (12:40 +0200)]
dt-bindings: usb: xhci-tegra: Add Tegra186 support

Extend the bindings to cover the set of features found in Tegra186.

Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>