Hans de Goede [Tue, 30 Sep 2014 16:45:32 +0000 (18:45 +0200)]
sunxi: Fix gmac not working reliable on the Bananapi
In order for the gmac nic to work reliable on the Bananapi, we need to set
bits 10-12 GTXDC "GMAC Transmit Clock Delay Chain" of the GMAC clk register
(0x01c20164) to 3.
Without this about 9 out of 10 ethernet packets get lost, with this setting
there is no packet loss.
So far setting these bits is only necessary on the Bananapi, so this commit
solves this with a bit of #ifdef CONFIG_BANANAPI code. If in the future we
need to do something similar for other boards, we can create a specific
CONFIG_FOO option for this then.
Simon Glass [Wed, 8 Oct 2014 04:01:49 +0000 (22:01 -0600)]
config: Move smdkv310 to use common exynos4 file
Most of the smdkv310 features are common with other exynos4 boards. To
permit easier addition of driver model support, use the common file and
add a device tree file.
Simon Glass [Wed, 8 Oct 2014 04:01:47 +0000 (22:01 -0600)]
exynos: config: Move cros_ec and tps65090 out of smdk boards
These boards do not in fact have a Chrome OS EC, nor a TPS565090 PMIC, so
move the settings into a separate common file to be used by those that need
it.
Simon Glass [Wed, 8 Oct 2014 04:01:45 +0000 (22:01 -0600)]
exynos: Move common exynos settings into a common file
Since exynos4 and exyno5 share many settings, we should move these into
a common file to avoid duplication.
In effect the changes are that all exynos boards now have EXT4 and FAT
write support. This affects exynos5250 and exynos5420 which previously
did not. This also disables the ext2 commands which are equivalent to
ext4 anyway.
Simon Glass [Wed, 8 Oct 2014 04:01:44 +0000 (22:01 -0600)]
exynos: Rename -dt config files to -common
We want exynos5250-dt.h to be a board which can support any exynos5250
device. This matches the naming used by Linux. As a first step, rename
the existing -dt files to -common to make it clear they are common files,
and not specific boards.
Simon Glass [Wed, 8 Oct 2014 04:01:42 +0000 (22:01 -0600)]
dm: exynos: Split out the cros_ec drivers
With the driver model conversion we are going to be using driver model for
SPI and not for I2C. This works OK so long as a board doesn't need both
dm and non-dm versions of the cros_ec driver. Since pit uses SPI and snow
uses I2C we need to split the configs so that only one driver is compiled
for each platform.
We can fix this later when driver model supports I2C.
Simon Glass [Wed, 8 Oct 2014 04:01:41 +0000 (22:01 -0600)]
cros_ec: exynos: Use the correct tps65090 driver in each case
Exynos 5250 boards (snow, spring) use the I2C driver but Exynos 5420 boards
cannot due to a hardware design decision. Select the correct driver to use
in each case.
Simon Glass [Wed, 8 Oct 2014 04:01:40 +0000 (22:01 -0600)]
cros_ec: power: Add a tunnelled version of the tps65090 driver
Unfortunately on Pit the AP has no direct access to the tps65090 but must
talk through the EC (over SPI) to the EC's I2C bus.
When driver model supports PMICs this will be relatively easy. In the
meantime the best approach is to duplicate the driver. It will be refactored
once driver model support is expanded.
Simon Glass [Wed, 8 Oct 2014 04:01:38 +0000 (22:01 -0600)]
Exynos: Use 900MHz ARM frequency in SPL for peach_pit
The device seems to hang in SPL if the full speed is used when booting from
USB, perhaps because the PMIC has not been set to the maximum ARM core
voltage yet. Slow it down to a reliable speed.
Eric Nelson [Wed, 1 Oct 2014 21:30:56 +0000 (14:30 -0700)]
usb: gadget: fastboot: terminate commands with NULL
Without NULL termination, various commands will read past the
end of input. In particular, this was noticed with error()
calls in cb_getvar and simple_strtoul() in cb_download.
Since the download callback happens elsewhere, the 4k buffer
should always be sufficient to handle command arguments.
Daniel Mack [Tue, 22 Jul 2014 10:47:02 +0000 (12:47 +0200)]
usb: musb-new: core: set MUSB_POWER_HSENAB in MUSB_POWER for host mode
This bit allows the MUSB controller to negotiate for high-speed mode when
the device is reset by the hub. If unset, Babble errors occur with
high-speed mass storage devices right after the first packet. This condition
is not caught by the interrupt handles in U-Boot, so no recovery is done,
and the USB communication is stuck.
To fix this, set the bit unconditionally, not only for
CONFIG_USB_GADGET_DUALSPEED but also for host-only modes.
Eric Nelson [Thu, 2 Oct 2014 19:16:52 +0000 (12:16 -0700)]
nitrogen6x: config: enable Android fastboot
Enable 'fastboot' command.
This is currently enabled but not yet functional. Including it in the
configuration will ease further testing and development as discussed
on the mailing list.
Troy Kisky [Thu, 2 Oct 2014 19:16:48 +0000 (12:16 -0700)]
nitrogen6x: config: allow more bootargs parameters
Increase the maximum number of arguments allowed by the Hush parser.
This prevents errors when users or scripts aren't quoting parameters
when setting the "bootargs" variable et al.
Eric Nelson [Thu, 2 Oct 2014 19:16:39 +0000 (12:16 -0700)]
nitrogen6x: display use I2C detect for HDMI
The HPD pin and RX_SENSE registers have proven to be less reliable
than using I2C on the EDID pins for detection of an HDMI monitor.
In particular, when the HDMI output is reset through a "reboot"
cycle, the detect_hdmi() routine often bounces, resulting in
a failure to detect a connected monitor.
Eric Nelson [Thu, 2 Oct 2014 19:16:33 +0000 (12:16 -0700)]
nitrogen6x: display: add support for LG-9.7 LVDS display
Add support for LG 9.7" LVDS panel (1024x768) with integrated eGalax
touch screen.
Note that this panel differs only slightly from the Hannstar XGA panel
(margins).
No auto-detection is available because it shares the same touch controller
as the Hannstar-XGA display, so you'll need to configure it through the
'panel' environment variable:
Eric Nelson [Thu, 2 Oct 2014 19:16:32 +0000 (12:16 -0700)]
nitrogen6x: display: add qvga panel
Add support for a 1/4 VGA panel with a 24-bit RGB interface.
No auto-detection is enabled, so you must configure the 'panel'
environment variable to use this display:
Robert Winkler [Thu, 2 Oct 2014 19:16:31 +0000 (12:16 -0700)]
nitrogen6x: display: add support lvds jeida screen
Add support for Boundary Devices 7" and 10.1" 1280x800 displays with
integrated FocalTech ft5x06 10-point touch controller.
Because they share the touch controller with the 1024x600 displays,
auto-detection is disabled and you must explicitly set the 'panel'
environment variable:
Eric Nelson [Thu, 2 Oct 2014 19:16:27 +0000 (12:16 -0700)]
nitrogen6x: staticize board file
Declare locally-used data structures and functions as
static and pull in header files to prevent compiler warnings
of "Should it be static?" when building with "make C=1".
Eric Nelson [Thu, 2 Oct 2014 19:16:22 +0000 (12:16 -0700)]
nitrogen6x: implement board_cfb_skip() to disable text output
Several customers have asked to leave the display quiet during
boot, so allow the user to express this request by the presence
of environment variable "novideo".
Pavel Machek [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: Split SoCFPGA configuration
Split the SoCFPGA configuration into SoC-specific part which is
common for all boards (socfpga_cyclone5_common.h) and a board
specific part. There is currently only one board, which is the
generic SoCFPGA board (socfpga_cyclone5.h), but there are more
to come.
This is necessary due to various features of the boards, which
unfortunatelly cannot be autodetected.
Marek Vasut [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: Clean up SoCFPGA configuration
Reorganize and cleanup the configuration file for SoCFPGA. There
is no functional change after this cleanup. This was necessary,
since the file was a wild mess and it was impossible to make sense
of it's content, let alone change something without breaking some
other thing. This patch puts the contents on par with regular U-Boot
standards.
Also remove unused preprocessor symbols CONFIG_SINGLE_BOOTOADER
and CONFIG_USE_IRQ, which is undefined by default. Finally, do
logical reordering of the defines in the file so it's much more
readable. The reordering was also necessary for the splitting
as the initial one was messy.
Marek Vasut [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: Add command to control HPS-FPGA bridges
Add command to enable and disable the bridges between HPS and FPGA.
This patch does have a checkpatch issue with the assembler portion,
checkpatch correctly complains that there should be no whitespace
before quoted newline. I do not agree that fixing this specific
checkpatch issue will improve the readability, thus this one is not
addressed.
Marek Vasut [Sun, 21 Sep 2014 11:57:40 +0000 (13:57 +0200)]
arm: socfpga: Move cache_enable to CPU code
Move icache_enable() and dcache_enable() function calls from
board code into the CPU code and into the enable_caches()
function. This is how the cache enabling code was designed
to work.
Add code which configures the AMBA NIC-301 and the SCU on the SoCFPGA .
The code sets the access permissions for the CPU to the AMBA slaves such
that the CPU can access them in both secure and non-secure mode.
Marek Vasut [Mon, 15 Sep 2014 01:58:22 +0000 (03:58 +0200)]
arm: socfpga: pl310: Map SDRAM to 0x0
Configure the PL310 address filter to make sure DRAM is mapped to 0x0.
This code also configures the "remap" register of NIC-301 and sets the
required 'mpuzero' bit.
Marek Vasut [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: reset: Add function to reset FPGA bridges
Add function to enable and disable FPGA bridges. This code is used
by the FPGA manager to disable the bridges before programming the
FPGA and will later be also used by the initialization code for the
chip to put the chip into well defined state during startup.
Add code necessary to program the FPGA part of SoCFPGA from U-Boot
with an RBF blob. This patch also integrates the code into the
FPGA driver framework in U-Boot so it can be used via the 'fpga'
command.
Add function to initialize the EMAC blocks upon board startup.
The preprocessor guards against building on SoCFPGA-VT and against
SPL build are not needed as those are handled implicitly via both
SPL framework and the socfpga_cyclone5.h config file, which will
not define CONFIG_DESIGNWARE_ETH if building for SoCFPGA-VT.
We cannot handle two EMAC ethernet blocks yet, therefore the ifdefs.
Once there is hardware using both EMAC blocks, this ifdef will have
to go.
Marek Vasut [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: reset: Add EMAC reset functions
Add functions to reset the EMAC ethernet blocks. We cannot handle
two EMAC ethernet blocks yet, therefore the ifdefs. Once there is
hardware using both EMAC blocks, this ifdef will have to go.
Marek Vasut [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: timer: Pull the timer reload value from config file
The timer reload value is a property of the timer hardware and there
is no reason for this to be configurable. Place this into the timer
driver just like on the other hardware.
Marek Vasut [Tue, 16 Sep 2014 17:54:32 +0000 (19:54 +0200)]
arm: socfpga: clock: Sync with reference code
Add the missing pieces from the reference clock code from Altera. This
puts the code on par with the Altera U-Boot fork for all but the SDRAM
self-refresh bits, which are not part of this patch.
Pavel Machek [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: clock: Add code to read clock configuration
Add the entire bulk of code to read out clock configuration from the SoCFPGA
CPU registers. This is important for MMC, QSPI and UART drivers as otherwise
they cannot determine the frequency of their upstream clock.
Marek Vasut [Mon, 8 Sep 2014 12:08:45 +0000 (14:08 +0200)]
arm: socfpga: clock: Drop nonsense inlining from clock manager code
The inlining is done by GCC when needed, there is no need to do it
explicitly. Furthermore, the inline keyword does not force-inline
the code, but is only a hint for the compiler. Scrub this hint.
Marek Vasut [Mon, 6 Oct 2014 15:45:55 +0000 (17:45 +0200)]
Merge branches 'topic/drivers/fpga-20141006', 'topic/drivers/mmc-20141006', 'topic/drivers/net-20141006', 'topic/tools/mkimage-20141006' and 'topic/arm/cache-20141006' into HEAD
projects / J-u-boot.git / log