Stefan Roese [Mon, 24 Apr 2017 07:48:02 +0000 (09:48 +0200)]
dm: core: Add DM_FLAG_OS_PREPARE flag
This new flag can be added to DM device drivers, which need to do some
final configuration before U-Boot exits and the OS (e.g. Linux) is
started. The remove functions of those drivers will get called at
this stage to do these last-stage configuration steps.
Signed-off-by: Stefan Roese <sr@denx.de> Reviewed-by: Simon Glass <sjg@chromium.org> Cc: Bin Meng <bmeng.cn@gmail.com>
Stefan Roese [Mon, 24 Apr 2017 07:48:01 +0000 (09:48 +0200)]
serial: serial-uclass: Use force parameter in stdio_deregister_dev()
On my x86 platform I've noticed, that calling dm_uninit() or the new
function dm_remove_devices_flags() does not remove the desired device at
all. Debugging showed, that the serial uclass returns -EPERM in
serial_pre_remove(). This patch sets the force parameter when calling
stdio_deregister_dev() resulting in a removal of the device.
Signed-off-by: Stefan Roese <sr@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Bin Meng [Fri, 21 Apr 2017 14:24:47 +0000 (07:24 -0700)]
x86: acpi: Fix Windows S3 resume failure
U-Boot sets up the real mode interrupt handler stubs starting from
address 0x1000. In most cases, the first 640K (0x00000 - 0x9ffff)
system memory is reported as system RAM in E820 table to the OS.
(see install_e820_map() implementation for each platform). So OS
can use these memories whatever it wants.
If U-Boot is in an S3 resume path, care must be taken not to corrupt
these memorie otherwise OS data gets lost. Testing shows that, on
Microsoft Windows 10 on Intel Baytrail its wake up vector happens to
be installed at the same address 0x1000. While on Linux its wake up
vector does not overlap this memory range, but after resume kernel
checks low memory range per config option CONFIG_X86_RESERVE_LOW
which is 64K by default to see whether a memory corruption occurs
during the suspend/resume (it's harmless, but warnings are shown
in the kernel dmesg logs).
We cannot simply mark the these memory as reserved in E820 table
because such configuration makes GRUB complain: unable to allocate
real mode page. Hence we choose to back up these memories to the
place where we reserved on our stack for our S3 resume work.
Before jumping to OS wake up vector, we need restore the original
content there.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Stefan Roese <sr@denx.de>
Bin Meng [Fri, 21 Apr 2017 14:24:44 +0000 (07:24 -0700)]
x86: acpi: Refactor acpi_resume()
To do something more in acpi_resume() like turning on ACPI mode,
we need locate ACPI FADT table pointer first. But currently this
is done in acpi_find_wakeup_vector().
This changes acpi_resume() signature to accept ACPI FADT pointer
as the parameter. A new API acpi_find_fadt() is introduced, and
acpi_find_wakeup_vector() is updated to use FADT pointer as the
parameter as well.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Stefan Roese <sr@denx.de>
Bin Meng [Fri, 21 Apr 2017 14:24:42 +0000 (07:24 -0700)]
x86: apci: Change PM1_CNT register access to RMW
In enter_acpi_mode() PM1_CNT register is changed to PM1_CNT_SCI_EN
directly without preserving its previous value. Update to change
the register access to read-modify-write (RMW).
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Stefan Roese <sr@denx.de>
Bin Meng [Fri, 21 Apr 2017 14:24:41 +0000 (07:24 -0700)]
x86: Adjust board_final_cleanup() order
Call board_final_cleanup() before write_tables(), so that anything
done in board_final_cleanup() on a normal boot path is also done
on an S3 resume path.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Stefan Roese <sr@denx.de>
Bin Meng [Fri, 21 Apr 2017 14:24:40 +0000 (07:24 -0700)]
x86: Do not clear high table area for S3
When SeaBIOS is being used, U-Boot reserves a memory area to be
used for configuration tables like ACPI. But it should not be
cleared otherwise ACPI table will be missing.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Stefan Roese <sr@denx.de>
Bin Meng [Fri, 21 Apr 2017 14:24:34 +0000 (07:24 -0700)]
x86: fsp: Mark memory used by U-Boot as reserved in the E820 table for S3
U-Boot itself as well as everything that is consumed by U-Boot (like
heap, stack, dtb, etc) needs to be reserved and reported in the E820
table when S3 resume is on.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Stefan Roese <sr@denx.de>
Philipp Tomsich [Fri, 5 May 2017 19:48:30 +0000 (21:48 +0200)]
video: bmp: rename CONFIG_BMP_24BMP to CONFIG_BMP_24BPP
Due to a typo, the 24 bit-per-pixel configuration ends in 24BMP
instead of 24BPP. This change renames it throughout the source tree
for consistency and to make moving these options into Kconfig easier
and less error-prone.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Hannes Schmelzer <hannes.schmelzer@br-automation.com>
Philipp Tomsich [Fri, 5 May 2017 19:48:26 +0000 (21:48 +0200)]
rockchip: video: introduce VIDEO_DW_HDMI and select for Rockchip HDMI
Instead of having drivers/video/rockchip/Kconfig point outside of its
hierarchy for dw_hdmi.o, we should use a configuration-option to
include the Designware HDMI support.
This change introduces a new config option (not to be selected via
menuconfig, but to be selected from a dependent video driver's
configuration option) that enables dw_hdmi.o and selects it whenever
the HDMI support for Rockchip SoCs is selected.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Some DVI monitors don't show anything in HDMI mode since audio stream
confuses them. To solve this situation, this commit adds HDMI flag in
timing data and sets it accordingly during edid parsing.
First existence of extension block is checked. If it exists and it is
CEA861 extension, then data blocks are checked for presence of HDMI
vendor specific data block. If it is present, HDMI flag is set.
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net> Reviewed-by: Simon Glass <sjg@chromium.org>
Masahiro Yamada [Mon, 15 May 2017 07:07:33 +0000 (16:07 +0900)]
kbuild: update DTC warning settings for bus and node/property name checks
Recent commits of DTC introduced new warnings checking PCI and simple
buses, unit address formatting, and stricter node and property name
checking. Disable the new DTC warnings by default. As before,
warnings are enabled with W=*. The strict node and property name
checks are a bit subjective, so they are only enabled for W=2.
(This policy reflects the commit 8654cb8d0371 of Linux.)
Tom Rini [Sat, 13 May 2017 02:33:28 +0000 (22:33 -0400)]
Kconfig: USB: Migrate CONFIG_USB_EHCI_HCD users to Kconfig
Migrate the rest of the users of CONFIG_USB_EHCI_HCD over to Kconfig.
For a few SoCs, imply or default y this if USB is enabled. In some
cases we had not already migrated to CONFIG_USB so do that as well.
Cc: Marek Vasut <marex@denx.de> Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Marek Vasut <marex@denx.de>
Tom Rini [Sat, 13 May 2017 02:33:25 +0000 (22:33 -0400)]
omap: Drop CONFIG_OMAP_VC_I2C_HS_MCODE
The symbol CONFIG_OMAP_VC_I2C_HS_MCODE always uses the default value.
Restructure the comment and code such that if a need arises later to use
another value we can address this then.
Tom Rini [Sat, 13 May 2017 02:33:24 +0000 (22:33 -0400)]
watchdog: Migrate OMAP_WATCHDOG to Kconfig
Move this entry to Kconfig. As it is a hardware watchdog, select
HW_WATCHDOG. While we could default to enabling this for all platforms,
it is currently only enabled by default on AM33XX, so keep that logic
today.
Cc: Roger Meier <r.meier@siemens.com> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Sat, 13 May 2017 02:33:23 +0000 (22:33 -0400)]
omap: spi: Drop CONFIG_OMAP3_SPI_D0_D1_SWAPPED support
This particular quirk is not enabled in any config files today. It does
however exist and is handled correctly in device trees and via
CONFIG_DM_SPI. So we drop the symbol now and add a comment to indicate
that any (new) boards that require this quirk need to enable DM_SPI
instead.
Tom Rini [Sat, 13 May 2017 02:33:22 +0000 (22:33 -0400)]
omap3: Migrate CONFIG_OMAP3_GPIO_X to Kconfig
The symbols CONFIG_OMAP3_GPIO_X control if we enable the clocks for a
given GPIO bank in U-Boot. select the required banks for each target.
In some cases we need to also migrate from CONFIG_USB_EHCI (deprecated,
in include/configs/) to CONFIG_USB_EHCI_HCD as we only require the GPIO
bank to be enabled if USB is also enabled.
Tom Rini [Sat, 13 May 2017 02:33:21 +0000 (22:33 -0400)]
gpio: Move OMAP_GPIO to Kconfig
This driver is used often enough such that we want to have this enabled
by default on any ARCH_OMAP2PLUS board, and this only compiles on
ARCH_OMAP2PLUS due to required defines, so mark that as the depends.
Tom Rini [Sat, 13 May 2017 02:33:19 +0000 (22:33 -0400)]
omap4: Drop redundant CONFIG_OMAP4430 symbol
While there are a few different OMAP4 SoCs, today we always set
CONFIG_OMAP4430 and CONFIG_OMAP44XX. Convert the few test of
CONFIG_OMAP4430 to CONFIG_OMAP44XX.
Cc: Marek Vasut <marex@denx.de> Cc: Paul Kocialkowski <contact@paulk.fr> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Sat, 13 May 2017 02:33:18 +0000 (22:33 -0400)]
omap3: Drop CONFIG_OMAP3_EVM, switch to CONFIG_TARGET_OMAP3_EVM when needed
We make use of CONFIG_OMAP3_EVM today to know when to do a specific
tweak in MUSB. This can be tested on via CONFIG_TARGET_OMAP3_EVM
instead, so switch there so we can drop the now unused symbol
CONFIG_OMAP3_EVM. In investigating what to do about the symbol usage we
see that the cairo board defines the same function, but never called it
(as it does not define CONFIG_OMAP3_EVM) and was just returning anyhow,
so drop that function from that board.
Cc: "Albert ARIBAUD (3ADEV)" <albert.aribaud@3adev.fr> Cc: Marek Vasut <marex@denx.de> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Sat, 13 May 2017 02:33:17 +0000 (22:33 -0400)]
omap5: Migrate CONFIG_OMAP_PLATFORM_RESET_TIME_MAX_USEC to Kconfig
While in theory this value could be used in places outside of "omap5"
(such as OMAP4), we only make use of it today in OMAP5, so place the
Kconfig entry there. Given that Kconfig lets us provide a default, we
drop CONFIG_DEFAULT_OMAP_RESET_TIME_MAX_USEC entirely. The contents of
doc/README.omap-reset-time make a good help entry, so adjust them
slightly and delete the file. Move the comment about range to where we
use the value now, and have Kconfig enforce the upper bound.
Tom Rini [Sat, 13 May 2017 02:33:16 +0000 (22:33 -0400)]
TI: Drop 'CONFIG_OMAP'
In the two cases in the code where we use CONFIG_OMAP as a useful test
currently we can make use of CONFIG_ARCH_OMAP2PLUS instead. With that
changed we can drop all defines of CONFIG_OMAP. While in here,
CONFIG_OMAP3430 is only defined and then never used, so drop.
Kever Yang [Fri, 5 May 2017 03:47:45 +0000 (11:47 +0800)]
spl: add support to booting with ATF
ATF(ARM Trusted Firmware) is used by ARM arch64 SoCs, find more infomation
about ATF at: https://github.com/ARM-software/arm-trusted-firmware
SPL is considered as BL2 in ATF terminology, it needs to load other parts
of ATF binary like BL31, BL32, SCP-BL30, and BL33(U-Boot). And needs to
prepare the parameter for BL31 which including entry and image information
for all other images. Then the SPL handle PC to BL31 with the parameter,
the BL31 will do the rest of work and at last get into U-Boot(BL33).
This patch needs work with patches from Andre for SPL support multi
binary in FIT.
The entry point of bl31 and bl33 are still using hard code because we
still can not get them from the FIT image information.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Tested-by: Heiko Stuebner <heiko@sntech.de> Acked-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
Masahiro Yamada [Tue, 9 May 2017 11:31:38 +0000 (20:31 +0900)]
blanche_defconfig: enable CONFIG_MMC
As commit 54925327fa11 ("mmc: move CONFIG_GENERIC_MMC to Kconfig")
addressed, this is one of the last weird defconfigs that define
CONFIG_GENERIC_MMC without CONFIG_MMC.
Now I took a closer look at this. Given that both CONFIG_CMD_MMC
and CONFIG_GENERIC_MMC are set for this defconfig, CONFIG_MMC should
be enabled.
As commit 54925327fa11 ("mmc: move CONFIG_GENERIC_MMC to Kconfig")
addressed, this is one of the last weird defconfigs that define
CONFIG_GENERIC_MMC without CONFIG_MMC.
Now I took a closer look at this. Given that neither CONFIG_CMD_MMC
nor CONFIG_MMC is set for this defconfig, CONFIG_GENERIC_MMC
should be disabled.
Masahiro Yamada [Tue, 9 May 2017 06:52:04 +0000 (15:52 +0900)]
mmc: sdhci-cadence: import updates from Linux 4.12
This driver is a counterpart of drivers/mmc/host/sdhci-cadence.c
from Linux. Some updates for v4.12-rc1 can be imported to U-Boot.
- Fix value of SDHCI_CDNS_HRS04_RDATA_SHIFT
- Add polling for ACK bit to be sure that data are written to
the PHY register
- Retrieve PHY values from DT properties instead of fixed data
Wenyou Yang [Wed, 26 Apr 2017 01:32:30 +0000 (09:32 +0800)]
mmc: sdhci: Fix maximum clock for programmable clock mode
In the programmable clock mode, the SDCLK frequency is incorrectly
assigned when the maximum clock has been assigned during probe,
this causes the SDHCI not work well.
In the programmable clock mode, when calculating the SDCLK Frequency
Select, when the maximum clock has been assigned, it is the actual
value, should not be multiplied by host->clk_mul. Otherwise, the
maximum clock is multiplied host->clk_mul by the base clock achieved
from the BASECLKF field of the Capabilities 0 Register.
Signed-off-by: Wenyou Yang <wenyou.yang@atmel.com>
Sekhar Nori [Thu, 11 May 2017 14:44:22 +0000 (20:14 +0530)]
davinci: omapl138_lcdk: switch to using TI_COMMON_CMD_OPTIONS
Now that we support using TI_COMMON_CMD_OPTIONS,
we dont have to enable a number of commands explicitly
in the defconfig if we enable TI_COMMON_CMD_OPTIONS.
Enable TI_COMMON_CMD_OPTIONS and regenerate the defconfig
using "make savedefconfig".
CMD_GPIO has been kept disabled because there is no GPIO
support (yet) on OMAP-L138 LCDK.
Signed-off-by: Sekhar Nori <nsekhar@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Tom Rini [Wed, 10 May 2017 19:26:06 +0000 (15:26 -0400)]
net: Disable the format-extra-args warning
We will see warnings such as:
net/eth_common.c:57:61: warning: data argument not used by format string [-Wformat-extra-args]
sprintf(enetvar, index ? "%s%daddr" : "%saddr", base_name, index);
~~~~~~~~ ^
With clang. In this case we do not want to re-write our code to be less
compact as the above is intentional and readable. Add a comment above
the disabling so that it's clear why we want that warning off.
Tom Rini [Wed, 10 May 2017 19:20:15 +0000 (15:20 -0400)]
gpio-uclass.c: Fix comparison of unsigned expression warning
We declare that gpio_base (which is the base for counting gpios, not an
address) is unsigned. Therefore the comparison with >= 0 is always
true. As the desire is to allow for this base number to be 0, we can
just drop this check. Reported by clang-3.8.
Cc: Simon Glass <sjg@chromium.org> Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Tom Rini [Wed, 10 May 2017 19:20:14 +0000 (15:20 -0400)]
cmd/led.c: Remove unnecessary check on 'cmd' value
We first check that if argc is less than 2 we return CMD_RET_USAGE. We
then see if argc is greater than 2 and if so call get_led_cmd() to set
'cmd' and otherwise set it to LEDST_COUNT (which will always be positive
as it's an enum with 0 already assigned). Therefore the test on if cmd
is less than 0 will always be false and simply be omitted. Reported by
clang-3.8.
Tom Rini [Wed, 10 May 2017 19:20:13 +0000 (15:20 -0400)]
cmd/io.c: Fix comparison of unsigned expression warning
The function cmd_get_data_size() returns an int and not unsigned. So we
should assign it to an int rather than unsigned so that we can later
compare the return value. Reported by clang-3.8.
Lokesh Vutla [Wed, 10 May 2017 10:49:52 +0000 (16:19 +0530)]
tools/genboardscfg.py: Make 'Supported' as known status
As per MAINTAINERS[1] file description, 'Supported' is
a valid status for a board. But buildman thinks 'Maintained'
is the only valid state and complains about boards with 'Supported'
status. Update buildman to accept 'Supported' as valid state.
Peng Fan [Tue, 9 May 2017 02:32:03 +0000 (10:32 +0800)]
asm-generic: global_data: change timebase_l/h to unsigned int
Change type of timebase_l/h to unsigned int.
>From lib/time.c: ((uint64_t)gd->timebase_h << 32) | gd->timebase_l;
This piece code is based on that timebase_h and timebase_l are
32bits width, so change the type to unsigned int.
Signed-off-by: Peng Fan <peng.fan@nxp.com> Cc: Eddie Cai <eddie.cai.linux@gmail.com> Cc: Jagan Teki <jteki@openedev.com> Cc: York Sun <york.sun@nxp.com> Cc: "Robert P. J. Day" <rpjday@crashcourse.ca> Cc: Michal Simek <michal.simek@xilinx.com> Cc: Tom Rini <trini@konsulko.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Peng Fan [Tue, 9 May 2017 02:32:02 +0000 (10:32 +0800)]
arm: change tbu/l type to unsigned int
Change tbu/l type to unsigned int.
>From the timer file for arm,
"(((unsigned long long)gd->arch.tbu) << 32) | gd->arch.tbl;" is used,
This piece code is based on tbu/tbl is 32bits, so change the type to
unsigned int.
Signed-off-by: Peng Fan <peng.fan@nxp.com> Cc: Simon Glass <sjg@chromium.org> Cc: Tom Rini <trini@konsulko.com>
Tom Rini [Tue, 9 May 2017 02:14:35 +0000 (22:14 -0400)]
m5253demo: Fix static variable in non-static inline function warning
The function 'spin_wheel' is declared as inline, but not static and thus
we see warnings that 'w' and 'p' are declared static in a non-static
inline function. Correct this by marking spin_wheel as static inline.
Cc: TsiChung Liew <Tsi-Chung.Liew@freescale.com> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Tue, 9 May 2017 02:14:34 +0000 (22:14 -0400)]
common: Only build cli_readline.o for CMDLINE on non-SPL
With gcc-6 and later we may get a warning such as:
.../common/cli_readline.c:20:21: warning: ‘tab_seq’ defined but not used [-Wunused-const-variable=]
static const char tab_seq[] = " "; /* used to expand TABs */
^~~~~~~
.../common/cli_readline.c:19:19: warning: ‘erase_seq’ defined but not used [-Wunused-const-variable=]
static const char erase_seq[] = "\b \b"; /* erase sequence */
^~~~~~~~~
Because in SPL we're normally not doing interactive commands anyhow, so
lets just not compile this at all in SPL. This also means that we need
to correct the logic (and comment) about what the drivers/ddr/fsl/ and
CONFIG_FSL_DDR_INTERACTIVE requires and this will be included in SPL
there.
Tom Rini [Tue, 9 May 2017 02:14:33 +0000 (22:14 -0400)]
net: uli526x: Fix unknown storage size error
The variable netdev_ethtool_ops is not referenced, drop it. However
with gcc-6 or later we fail to even compile as we do not have the
required struct definition in U-Boot.
Cc: Joe Hershberger <joe.hershberger@ni.com> Signed-off-by: Tom Rini <trini@konsulko.com>
The variable 'res' may be unused uninitialized if our call to
mv88e61xx_port_read (register read) fails and we goto the error
handling section. In this case we set 'res' to -EIO to indicate why we
failed.
Cc: Joe Hershberger <joe.hershberger@ni.com> Cc: Chris Packham <judge.packham@gmail.com> Cc: Kevin Smith <kevin.smith@elecsyscorp.com> Cc: Prafulla Wadaskar <prafulla@marvell.com> Signed-off-by: Tom Rini <trini@konsulko.com>
projects / u-boot.git / log