Marek Vasut [Mon, 25 Jan 2016 21:00:44 +0000 (22:00 +0100)]
usb: kbd: Prevent out of bound access
Scan code 0x39 is CapsLock, which is not a printable character and thus
is not covered by either usb_kbd_numkey_shifted[] or usb_kbd_numkey[].
Fix the scan code check to avoid looking it up in either of the arrays.
Ted Chen [Wed, 20 Jan 2016 06:24:52 +0000 (14:24 +0800)]
usb: eth: add Realtek RTL8152B/RTL8153 DRIVER
This patch adds driver support for the Realtek RTL8152B/RTL8153 USB
network adapters.
Signed-off-by: Ted Chen <tedchen at realtek.com>
[swarren, fixed a few compiler warnings]
[swarren, with permission, converted license header to SPDX]
[swarren, removed printf() spew during probe()] Signed-off-by: Stephen Warren <swarren at nvidia.com>
Stephen Warren [Wed, 16 Dec 2015 03:50:22 +0000 (20:50 -0700)]
ARM: rpi: enable USB keyboard
Now that the DWC2 driver supports split transactions, we can reasonably
enable support for USB keyboards. This wasn't terribly useful before
since keyboards are usually LS/FS devices, and thus require split
transaction support when attached to a USB hub such as the hub built into
the RPi model Bs.
Cc: Stefan Brüns <stefan.bruens@rwth-aachen.de> Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Stefan Brüns [Sat, 23 Jan 2016 00:42:25 +0000 (01:42 +0100)]
usb: dwc2: Do not mix data toggle for IN and OUT endpoints, check bounds
USB protocol allows for 16 IN and 16 OUT endpoints (USB 2.0 Spec,
8.3.2.2 Endpoint Field). A function may have an EP 1 for both IN and OUT,
so these two should be kept separate. As EPs are either BULK or INTERRUPT
(or ISO), it is fine to have one array per direction for all transfer
types (also see e236519b7365ef75c5da6a5623f0b03d9c00cfae).
USB device address is 7 bits, so a bus may have more than 16 devices.
Check the device number, as the DWC2 driver only supports BULK/ISO for
the first 16 devices.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Stefan Brüns [Sun, 17 Jan 2016 03:09:56 +0000 (04:09 +0100)]
usb: dwc2: Add SPLIT INTERRUPT transaction support
CSPLITs for INTERRUPT transactions have to be scheduled in each microframe
following the SSPLIT. INTERRUPT transfers are executed in the next even/
odd microframe depending on the HCCHAR_ODDFRM flag.
As there are no handshakes for INTERRUPT SSPLITs the SSPLIT may have
failed (transport error) without the error being detected by the host
driver. If the last CSPLIT is not received within 4 microframes after the
SSPLIT there was a transaction error and the complete transaction has
to be restarted.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Stefan Brüns [Sun, 17 Jan 2016 03:09:55 +0000 (04:09 +0100)]
usb: dwc2: Implement SPLIT transaction support
In contrast to non-SPLIT transfers each transaction has to be submitted
as an individual chunk.
The transaction state machine proceeds from SSPLIT to CSPLIT if the ACK
flag is set. CSPLIT has to be repeated while NYET is set.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Stefan Brüns [Sun, 17 Jan 2016 03:09:53 +0000 (04:09 +0100)]
usb: dwc2: split transfer core from outer loop
Split the movement of data between CPU and Host Controller from the
status handling and tracking of transfer progress.
This will also simplify adding of SPLIT transaction support.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
A transfer is completed if the XFERCOMP flag is set, irrespective of the
ACK flag. BULK OUT transfers to some HS devices complete without having
the ACK flag set, which signal the devices has responded with an NYET
to the transfer (PING protocol).
The new behaviour matches the Linux kernel minus any PING protocol.
Also see 5966defabdcc (usb: dwc2: fix bulk transfers)
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Fix two errors in transfer len calculation, move loop invariant code out
of loop.
If xfer_len is equal to CONFIG_DWC2_MAX_TRANSFER_SIZE (or slightly
smaller), the xfer_len will be to large, e.g.:
xfer_len = MAX_TRANSFER_SIZE = 65535
max packet size = 512
=> num_packets = 128
=> IN xfer_len = 65536
For OUT transactions larger than (65536 - mps) bytes, the xfer_len
determination is quite awkward, it is only correct due to:
- max_packet_size for control/bulk/interrupt is required to be
power-of-two.
- (CONFIG_DWC2_MAX_TRANSFER_SIZE + 1) % max-packet-size is zero
for all allowed (2^3 ... 2^9) packet sizes
As the max xfer len is loop invariant, it can be moved out of the loop.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Simon Glass [Fri, 22 Jan 2016 02:45:25 +0000 (19:45 -0700)]
rockchip: Update the README
GPIO, I2C, LCD and HDMI are now implemented. We have more than one PMIC.
There is an implementation to run the CPU at full speed although it does
not seem to make much difference.
Simon Glass [Fri, 22 Jan 2016 02:45:19 +0000 (19:45 -0700)]
rockchip: jerry: Enable EDP and HDMI video output
Enable these devices using the VOPL video output device. We explicitly
disable VOPB in the device tree to avoid it taking over. Since this device
has an LCD display this comes up by default. If the display fails for some
reason then it will attempt to use HDMI. It is possible to force it to fail
(and thus fall back to HDMI) by puting 'return -EPERM' at the top of
rk_edp_probe(). For now there is no easy way to select between the two.
Simon Glass [Fri, 22 Jan 2016 02:45:08 +0000 (19:45 -0700)]
rockchip: Add a script to parse datasheets
This script has proved useful for parsing datasheets and creating register
shift/mask values for use in header files. Include it in case it is useful
for others.
Simon Glass [Fri, 22 Jan 2016 02:45:05 +0000 (19:45 -0700)]
rockchip: video: Add a video-output driver
Some rockchip SoCs include video output (VOP). Add a driver to support this.
It can output via a display driver (UCLASS_DISPLAY) and currently HDMI and
eDP are supported.
Simon Glass [Fri, 22 Jan 2016 02:45:03 +0000 (19:45 -0700)]
rockchip: video: Add a display driver for rockchip HDMI
Some Rockchip SoCs support HDMI output. Add a display driver for this so
that these displays can be used on supported boards.
Unfortunately this driver is not fully functional. It cannot reliably read
EDID information over HDMI. This seems to be due to the clocks being
incorrect - the I2C bus speed appears to be up to 100x slower than the
clock settings indicate. The root cause may be in the clock logic.
Simon Glass [Fri, 22 Jan 2016 02:45:00 +0000 (19:45 -0700)]
dm: video: Repurpose the 'displayport' uclass to 'display'
The current DisplayPort uclass is too specific. The operations it provides
are shared with other types of output devices, such as HDMI and LVDS LCD
displays.
Generalise the uclass so that it can be used with these devices as well.
Adjust the uclass to handle the EDID reading and conversion to
display_timing internally.
Simon Glass [Fri, 22 Jan 2016 02:44:59 +0000 (19:44 -0700)]
video: panel: Add a simple panel driver
Most panels are very simple - they just have a power supply and a backlight.
Add a driver which supports this and implements the enable_backlight()
method.
Simon Glass [Fri, 22 Jan 2016 02:44:58 +0000 (19:44 -0700)]
dm: panel: Add a panel uclass
LCD panels can usefully be modelled as their own uclass. They can be probed
(which powers them up ready for use). If they have a backlight, this can be
enabled.
Simon Glass [Fri, 22 Jan 2016 02:44:56 +0000 (19:44 -0700)]
dm: backlight: Add a backlight uclass
LCD panels normally have a backlight which can be controlled to illuminate
the LCD contents. Add a uclass to support this. Initially it only has a
method to enable the backlight.
Simon Glass [Fri, 22 Jan 2016 02:44:51 +0000 (19:44 -0700)]
video: Name consoles by their number
We must use the console name in the 'stdout' variable to select the one
we want. At present the name is formed from the driver name with a suffix
indicating the rotation value.
It seems better to name them sequentially since this can be controlled by
driver order. So adjust the code to use 'vidconsole' for the first,
'vidconsole1' for the second, etc.
Simon Glass [Fri, 22 Jan 2016 02:44:50 +0000 (19:44 -0700)]
gpio: Warn about invalid GPIOs used with the 'gpio' command
At present there is no indication that an invalid GPIO is used except that
the GPIO status is not displayed. Make the error more explicit to avoid
confusion.
Simon Glass [Fri, 22 Jan 2016 02:44:12 +0000 (19:44 -0700)]
rockchip: spi: Remove the explicit pinctrl setting
The correct pinctrl is handled automatically so we don't need to do it in
the driver. The exception is when we want to use a different chip select
(other than 0). But this isn't used at present.
Simon Glass [Fri, 22 Jan 2016 02:44:11 +0000 (19:44 -0700)]
rockchip: spi: Correct chip-enable code
At present there is an incorrect call to rkspi_enable_chip(). It should
be disabling the chip, not enabling it. Correct this and ensure that the
chip is disabled when releasing the bus.
Simon Glass [Fri, 22 Jan 2016 02:44:05 +0000 (19:44 -0700)]
rockchip: clk: Make rkclk_get_clk() SoC-specific
The current method assumes that clocks are numbered from 0 and we can
determine a clock by its number. It is safer to use an ID in the clock's
platform data to avoid the situation where another clock is bound before
the one we expect.
Move the existing code into rk3036 since it still works there. Add a new
implementation for rk3288.
Simon Glass [Fri, 22 Jan 2016 02:43:59 +0000 (19:43 -0700)]
dm: power: Allow regulators to not implement all operations
Some regulators will not implement any operations (e.g. fixed regulators).
This is not an error, so allow the autoset process to continue when one
of these regulators is found.
Simon Glass [Fri, 22 Jan 2016 02:43:58 +0000 (19:43 -0700)]
dm: power: Tidy up debugging output and return values
The currect PMIC debugging is a little confusing. Adjust it so that it is
clear whether the operation succeeded or failed. Also, avoid creating a new
error return value when a perfectly good one is already available.
Simon Glass [Fri, 22 Jan 2016 02:43:56 +0000 (19:43 -0700)]
dm: pinctrl: Add a way for a GPIO driver to obtain a pin function
GPIO drivers want to be able to show if a pin is enabled for input, output,
or is being used by another function. Some drivers can easily find this
and the code is included in the driver. For some SoCs this is more complex.
Conceptually this should be handled by pinctrl rather than GPIO. Most
pinctrl drivers will have this feature anyway.
Add a method by which a GPIO driver can obtain the pin mux value given a
GPIO reference. This avoids repeating the code in two places.
Simon Glass [Fri, 22 Jan 2016 02:43:55 +0000 (19:43 -0700)]
dm: power: Allow regulators to be omitted from SPL
For some boards the pmic interface is useful but the regulator interface
(which comes with it) is too large. Allow them to be separated such that
SPL can decide which it needs.
Simon Glass [Fri, 22 Jan 2016 02:43:54 +0000 (19:43 -0700)]
spi: Correct device tree usage in spi_flash_decode_fdt()
This function currently searches the entire device tree for a node that
it thinks is relevant. But the node is known and is passed in. Correct the
code and enable it only with driver model, since only driver-model boards
will use it.
This avoids bringing in a large number of strings from fdtdec.
Simon Glass [Fri, 22 Jan 2016 02:43:44 +0000 (19:43 -0700)]
rockchip: spi: Avoid setting the pinctrl twice
If full pinctrl is enabled we don't need to manually set the pinctrl in the
driver. It will happen automatically. Adjust the code to suit - we will
still use manual mode in SPL.
Simon Glass [Fri, 22 Jan 2016 02:43:41 +0000 (19:43 -0700)]
rockchip: clock: Add a function to find a clock by ID
The current approach of using uclass_get_device() is error-prone. Another
clock (for example a fixed-clock) may cause it to break. Add a function that
does a proper search.
Simon Glass [Fri, 22 Jan 2016 02:43:38 +0000 (19:43 -0700)]
rockchip: Use a separate clock ID for clocks
At present we use the same peripheral ID for clocks and pinctrl. While this
works it is probably better to use the device tree clock binding ID for
clocks. We can use the clk_get_by_index() function to find this.
Update the clock drivers and the code that uses them.
Simon Glass [Fri, 22 Jan 2016 02:43:26 +0000 (19:43 -0700)]
dm: pinctrl: Add a function to parse PIN_CONFIG flags
Add a function which produces a flags word from a few common PIN_CONFIG
settings. This is useful for simple pinctrl drivers that don't need to worry
about drive strength, etc.
Simon Glass [Fri, 22 Jan 2016 02:43:25 +0000 (19:43 -0700)]
dm: core: Don't set pinctrl for pinctrl devices
There is sort-of race condition when a pinctrl device is probed. The pinctrl
function is called which may end up using the same device as is being
probed. This results in operations being used before the device is actually
probed.
For now, disallow pinctrl operations on pinctrl devices while probing. An
alternative solution would be to move the operation to later in the
device_probe() function (for pinctrl devices only) but this needs more
thought.