oweals/u-boot.git
5 years agoblk: Increase cache element size
Marek Vasut [Wed, 8 Aug 2018 11:20:29 +0000 (13:20 +0200)]
blk: Increase cache element size

Cache up to 4 kiB entries. 4 kiB is the default block size on ext4, yet
the underlying block layer devices usually report support for 512B . In
most cases, the 512B support is emulated (ie. SD cards, SSDs, USB sticks
etc.) and the real block size of those devices is much bigger.

To avoid performance degradation with such devices and FS setup, bump
the maximum cache entry size to 4 kiB.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Tom Rini <trini@konsulko.com>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
5 years agoMakefile: run CONFIG_BOARD_SIZE_LIMIT against .img
Simon Goldschmidt [Fri, 11 Jan 2019 19:56:02 +0000 (20:56 +0100)]
Makefile: run CONFIG_BOARD_SIZE_LIMIT against .img

With the current Makefile, CONFIG_BOARD_SIZE_LIMIT is used to check
the U-Boot binary without devicetree only. This produces wrong results
when OF_SEPARATE is used.

To fix this, run the CONFIG_BOARD_SIZE_LIMIT check on all .img binaries
as well.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agokbuild: add .SECONDARY special target to scripts/Kbuild.include
Masahiro Yamada [Fri, 11 Jan 2019 10:42:27 +0000 (19:42 +0900)]
kbuild: add .SECONDARY special target to scripts/Kbuild.include

Based on the following Linux commits:

 - 54a702f70589 ("kbuild: mark $(targets) as .SECONDARY and remove
   .PRECIOUS markers")

 - 8e9b61b293d9 ("kbuild: move .SECONDARY special target to
   Kbuild.include")

GNU Make automatically deletes intermediate files that are updated
in a chain of pattern rules.

Example 1) %.dtb.o <- %.dtb.S <- %.dtb <- %.dts
Example 2) %.o <- %.c <- %.c_shipped

A couple of makefiles mark such targets as .PRECIOUS to prevent Make
from deleting them, but the correct way is to use .SECONDARY.

  .SECONDARY
    Prerequisites of this special target are treated as intermediate
    files but are never automatically deleted.

  .PRECIOUS
    When make is interrupted during execution, it may delete the target
    file it is updating if the file was modified since make started.
    If you mark the file as precious, make will never delete the file
    if interrupted.

Both can avoid deletion of intermediate files, but the difference is
the behavior when Make is interrupted; .SECONDARY deletes the target,
but .PRECIOUS does not.

The use of .PRECIOUS is relatively rare since we do not want to keep
partially constructed (possibly corrupted) targets.

.SECONDARY with no prerequisites causes all targets to be treated as
secondary. This agrees the policy of Kbuild.

scripts/Kbuild.include seems a suitable place to add it because it is
included from almost all sub-makes.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agokbuild: add .DELETE_ON_ERROR special target
Masahiro Yamada [Fri, 11 Jan 2019 10:42:26 +0000 (19:42 +0900)]
kbuild: add .DELETE_ON_ERROR special target

Linux commit 9c2af1c7377a8a6ef86e5cabf80978f3dbbb25c0

If Make gets a fatal signal while a shell is executing, it may delete
the target file that the recipe was supposed to update.  This is needed
to make sure that it is remade from scratch when Make is next run; if
Make is interrupted after the recipe has begun to write the target file,
it results in an incomplete file whose time stamp is newer than that
of the prerequisites files.  Make automatically deletes the incomplete
file on interrupt unless the target is marked .PRECIOUS.

The situation is just the same as when the shell fails for some reasons.
Usually when a recipe line fails, if it has changed the target file at
all, the file is corrupted, or at least it is not completely updated.
Yet the file’s time stamp says that it is now up to date, so the next
time Make runs, it will not try to update that file.

However, Make does not cater to delete the incomplete target file in
this case.  We need to add .DELETE_ON_ERROR somewhere in the Makefile
to request it.

scripts/Kbuild.include seems a suitable place to add it because it is
included from almost all sub-makes.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agocommon: Kconfig: miscellaneous spelling fixes
Chris Packham [Fri, 11 Jan 2019 02:30:50 +0000 (15:30 +1300)]
common: Kconfig: miscellaneous spelling fixes

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agokbuild: fix parallel build race caused by u-boot.cfg regeneration
Masahiro Yamada [Thu, 10 Jan 2019 14:11:39 +0000 (23:11 +0900)]
kbuild: fix parallel build race caused by u-boot.cfg regeneration

Multiple people have reported intermittent build failure in parallel
building.

Kever Yang reported this issue some time ago [1], but I could not
get enough clue at that time.

This time, Richard Purdie provided a full build log [2], which was
very helpful for me to root-cause it.

The cause of the problem is commit 0d982c585330 ("Makefile: add
dependencies to regenerate u-boot.cfg when lost").

That commit added the 'cfg' as the prerequisite of the 'all' target,
so the parallel build tries to run it simultaneously, then regenerates
a symlink while building objects.

When u-boot.cfg is accidentally lost, let's rebuild it before
descending into any subdirectories.

Also, what is annoying is u-boot.cfg is currently regenerated every
time since it depends on FORCE. We can get rid of all the prerequisites
of u-boot.cfg because u-boot.cfg is rebuilt anyway as the byproduct of
auto.conf when a user updates the .config file.

[1] https://lists.denx.de/pipermail/u-boot/2018-June/330341.html
[2] https://autobuilder.yoctoproject.org/typhoon/#/builders/65/builds/160/steps/7/logs/step1b

Fixes: 0d982c585330 ("Makefile: add dependencies to regenerate u-boot.cfg when lost")
Reported-by: Kever Yang <kever.yang@rock-chips.com>
Reported-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoserial: ns16550: fix debug uart putc called before init
Simon Goldschmidt [Wed, 9 Jan 2019 19:35:31 +0000 (20:35 +0100)]
serial: ns16550: fix debug uart putc called before init

If _debug_uart_putc() is called before _debug_uart_init(), the
ns16550 debug uart driver hangs in a tight loop waiting for the
tx FIFO to get empty.

As this can happen via a printf sneaking in before the port calls
debug_uart_init(), introduce a config option to ignore characters
before the debug uart is initialized.

This is done by reading the baudrate divisor and aborting if is zero.

The Kconfig option is required as reading the baudrate divisor does
not seem to work for all ns16500 compatibles (which is why the last
attempt on this has been reverted in 1a67969a99).

Tested on socfpga_cyclone5_socrates.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agodrivers: serial: DEBUG_UART_SKIP_INIT depends on DEBUG_UART
Simon Goldschmidt [Wed, 9 Jan 2019 19:27:09 +0000 (20:27 +0100)]
drivers: serial: DEBUG_UART_SKIP_INIT depends on DEBUG_UART

DEBUG_UART_SKIP_INIT is used only by debug UART and thus should depend
on DEBUG_UART.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoFix typo: missmatched -> mismatched.
Vagrant Cascadian [Tue, 8 Jan 2019 21:10:23 +0000 (13:10 -0800)]
Fix typo: missmatched -> mismatched.

Signed-off-by: Vagrant Cascadian <vagrant@debian.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
5 years agoxyz-modem: Fix timeout loop waiting with WATCHDOG
Lokesh Vutla [Tue, 8 Jan 2019 13:58:35 +0000 (19:28 +0530)]
xyz-modem: Fix timeout loop waiting with WATCHDOG

Commit 2c77c0d6524eb ("xyz-modem: Change getc timeout loop waiting")
fixes the loop delay when using a hw watchdog, assuming that watchdog
kicking is taken care of by getc(). But the xyzmodem driver tries to
do a getc only after confirming that a character is available like below:
while (!tstc()) {
till timeout;
}
if (tstc())
*c = getc();

and getc() does a watchdog reset only if it fails to see a character.
In this case, getc() always sees a character and never does a
watchdog reset. So to make sure that watchdog doesn't get reset
while loading the file, do a watchdog reset just before starting the
image loading.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Vignesh R <vigneshr@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agopylibfdt: Use Python 2 in Makefile
Josef Lusticky [Tue, 8 Jan 2019 13:19:46 +0000 (14:19 +0100)]
pylibfdt: Use Python 2 in Makefile

pylibfdt needs Python 2 to build.
Replace $(PYTHON) with $(PYTHON2) in pylibfdt Makefile
to ensure Python 2 is used to build it.

This fixes build on systems where Python 3 is the default version
of the "python" interpreter.
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agolib: uuid: Do not enable UUID command SPL
Marek Vasut [Mon, 7 Jan 2019 20:23:38 +0000 (21:23 +0100)]
lib: uuid: Do not enable UUID command SPL

The uuid command is only really useful in U-Boot, but it's useless in
SPL. Worse yet, it pulls in various environment manipulation functions
as it call env_set(). Do not compile the command in in SPL.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Tom Rini <trini@konsulko.com>
5 years agospl: ymodem: Add support for loading gzip compressed uImage
Marek Vasut [Mon, 7 Jan 2019 20:23:22 +0000 (21:23 +0100)]
spl: ymodem: Add support for loading gzip compressed uImage

Add support for gunzip-ing gzip-compressed uImages in the SPL Ymodem code.
Loading data over Ymodem can be gruelingly slow, gzip-ing the data can
reduce that aggravating slowness at least slightly (depends on the data,
u-boot.bin compresses to ~1/3 of it's original size on ARM64), hence add
optional support for decompressing gzip-compressed uImages.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Tom Rini <trini@konsulko.com>
5 years agodtoc: make generated platdata structs const
Simon Goldschmidt [Mon, 7 Jan 2019 19:29:26 +0000 (20:29 +0100)]
dtoc: make generated platdata structs const

The platdata initialization structs are currently generated into .rwdata.
Make sure the are put into .rodata by generating them as const.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agocmd: zip: use correct format code
Heinrich Schuchardt [Sun, 6 Jan 2019 11:38:37 +0000 (12:38 +0100)]
cmd: zip: use correct format code

dst_len is defined as unsigned long. So use %lu for printf().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agocmd: unzip: use correct format code
Heinrich Schuchardt [Sun, 6 Jan 2019 11:34:16 +0000 (12:34 +0100)]
cmd: unzip: use correct format code

src_len is defined as unsigned long. So use %lu for printf().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agocmd: ubi: remove unreachable code
Heinrich Schuchardt [Sun, 6 Jan 2019 11:26:28 +0000 (12:26 +0100)]
cmd: ubi: remove unreachable code

It does not make sense to check if argc < 2 a second time, especially after
accessing argv[1].

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agocmd: tpm-v2: use correct format code
Heinrich Schuchardt [Sun, 6 Jan 2019 11:09:10 +0000 (12:09 +0100)]
cmd: tpm-v2: use correct format code

updates is defined as unsigned int. So use %u for printf().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agocmd: sf: use correct printf code
Heinrich Schuchardt [Sun, 6 Jan 2019 11:03:37 +0000 (12:03 +0100)]
cmd: sf: use correct printf code

test->time_ms[] is defined as unsigned. So use %u for printf().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agocmd: nvedit: use correct format code
Heinrich Schuchardt [Sun, 6 Jan 2019 10:52:06 +0000 (11:52 +0100)]
cmd: nvedit: use correct format code

len is defined as unsigned. So use %u for printf().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agocmd: gpio: use correct printf code
Heinrich Schuchardt [Sun, 6 Jan 2019 10:31:54 +0000 (11:31 +0100)]
cmd: gpio: use correct printf code

gpio is defined as unsigned int. So we should use %u when calling printf().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoconfigs: stm32: Remove CONFIG_OF_EMBED
Patrice Chotard [Fri, 4 Jan 2019 10:48:55 +0000 (11:48 +0100)]
configs: stm32: Remove CONFIG_OF_EMBED

Building with CONFIG_OF_EMBED generates build warnings, as it should
only be used for debugging purposes.

Remove CONFIG_OF_EMBED from all stm32 config files which are using
this flag.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
5 years agocmd: mmc: Invalidate MMC block cache after init
Marek Vasut [Thu, 3 Jan 2019 21:09:44 +0000 (22:09 +0100)]
cmd: mmc: Invalidate MMC block cache after init

Make sure the block cache is cleared for the MMC device after it was
reinitialized to avoid having any stale data in the cache, like e.g.
partition tables or such.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
5 years agocmd: mmc: Force mmc reinit when no card present
Marek Vasut [Thu, 3 Jan 2019 21:09:43 +0000 (22:09 +0100)]
cmd: mmc: Force mmc reinit when no card present

In case the card is removed, force-init the MMC to start the internal
machinery which deregisters and invalidate the MMC device.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
5 years agommc: Add support for downgrading HS200/HS400 to HS mode
Marek Vasut [Thu, 3 Jan 2019 20:19:24 +0000 (21:19 +0100)]
mmc: Add support for downgrading HS200/HS400 to HS mode

The mmc_select_mode_and_width() function can be called while the card
is in HS200/HS400 mode and can be used to downgrade the card to lower
mode, e.g. HS. This is used for example by mmc_boot_part_access_chk()
which cannot access the card in HS200/HS400 mode and which is in turn
called by saveenv if env is in the MMC.

In such case, forcing the card clock to legacy frequency cannot work.
Instead, the card must be switched to HS mode first, from which it can
then be reprogrammed as needed.

However, this procedure needs additional code changes, since the current
implementation checks whether the card correctly switched to HS mode in
mmc_set_card_speed(). The check only expects that the card will be going
to HS mode from lower modes, not from higher modes, hence add a parameter
which indicates that the HS200/HS400 to HS downgrade is happening. This
makes the code send the switch command first, reconfigure the controller
next and finally perform the EXT_CSD readback check. The last two steps
cannot be done in reverse order as the card is already in HS mode when
the clock are being switched on the controller side.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
5 years agoam335x: igep003x: Add Device Tree Support and DM_MMC driver
Enric Balletbo i Serra [Fri, 28 Dec 2018 10:55:48 +0000 (11:55 +0100)]
am335x: igep003x: Add Device Tree Support and DM_MMC driver

This adds device tree and the DM_MMC driver for the AM335x IGEP based
boards.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
5 years agoam335x: sl50: Add Device Tree Support and DM_MMC driver
Enric Balletbo i Serra [Thu, 27 Dec 2018 16:34:11 +0000 (17:34 +0100)]
am335x: sl50: Add Device Tree Support and DM_MMC driver

This adds device tree and the DM_MMC driver for the SL50 board.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
5 years agoomap3: igep00x0: Add Device Tree Support and DM_MMC driver
Enric Balletbo i Serra [Thu, 27 Dec 2018 16:03:30 +0000 (17:03 +0100)]
omap3: igep00x0: Add Device Tree Support and DM_MMC driver

This adds device tree for OMAP3 IGEP based boards and the DM_MMC driver.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
5 years agoomap3: igep00x0: Switch to simple malloc in SPL
Enric Balletbo i Serra [Thu, 27 Dec 2018 16:03:29 +0000 (17:03 +0100)]
omap3: igep00x0: Switch to simple malloc in SPL

To save more space, switch to simple malloc here.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
5 years agoomap3: igep00x0: Switch to using TI_COMMON_CMD_OPTION
Enric Balletbo i Serra [Thu, 27 Dec 2018 16:03:28 +0000 (17:03 +0100)]
omap3: igep00x0: Switch to using TI_COMMON_CMD_OPTION

Enable TI_COMMON_CMD_OPTIONS and remove similar options
from the defconfig.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
5 years agoomap3: igep00x0: Remove USB support due DM_USB deadline
Enric Balletbo i Serra [Thu, 27 Dec 2018 16:03:27 +0000 (17:03 +0100)]
omap3: igep00x0: Remove USB support due DM_USB deadline

The USB support for this board was never really tested, in fact, the
presence of these options are more a copy & paste error from the
Beagleboard than a feature that really was used. As doesn't work, remove
for now. If someone at some point want to add this support he'll need to
migrate the board to use CONFIG_DM_USB instead.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
5 years agoomap3: igep00x0: Remove unmaintained IGEP0032 defconfig
Enric Balletbo i Serra [Thu, 27 Dec 2018 16:03:26 +0000 (17:03 +0100)]
omap3: igep00x0: Remove unmaintained IGEP0032 defconfig

The IGEP0032 board was never officially pushed upstream and actually I
don't have access to this hardware, unless someone with the hardware
wants to start working on this doesn't makes sense have this defconfig
here. So remove it.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
5 years agotest: Use single quote consistently
Simon Glass [Thu, 27 Dec 2018 15:11:13 +0000 (08:11 -0700)]
test: Use single quote consistently

Some tests have ended up using double quotes where single quotes could be
used. Adjust this for consistency with the rest of U-Boot's Python code.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agoarm: dts: Add MMC nodes for HiKey board
Manivannan Sadhasivam [Thu, 27 Dec 2018 13:34:05 +0000 (19:04 +0530)]
arm: dts: Add MMC nodes for HiKey board

Add MMC nodes for HiKey board based on HI6220 SoC. There are three MMC
controllers in this SoC, first one used for eMMC, second one used
for SD card and third one is not used by u-boot.

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
5 years agommc: Convert HI6220 MMC driver to driver model
Manivannan Sadhasivam [Thu, 27 Dec 2018 13:34:04 +0000 (19:04 +0530)]
mmc: Convert HI6220 MMC driver to driver model

Convert HiSilicon HI6220 MMC driver based on DWMMC IP to driver
model.

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
[trini: Enable this on poplar]
Signed-off-by: Tom Rini <trini@konsulko.com>
5 years agoinclude: configs: Add gunzip size for HiKey board
Manivannan Sadhasivam [Thu, 27 Dec 2018 13:34:03 +0000 (19:04 +0530)]
include: configs: Add gunzip size for HiKey board

Default 8MB gunzip size is not enough to load the release kernel, hence
fix 64MB size for uncompressing the kernel.

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoconfigs: MediaTek: use OF_SEPARATE instead of OF_EMBED
Weijie Gao [Thu, 20 Dec 2018 08:12:59 +0000 (16:12 +0800)]
configs: MediaTek: use OF_SEPARATE instead of OF_EMBED

This patch replace OF_EMBED with OF_SEPARATE of defconfig files of
MediaTek boards because now OF_EMBED is only used for debugging purpose.

Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoMAINTAINERS: ARM MEDIATEK: update file entries
Weijie Gao [Thu, 20 Dec 2018 08:12:58 +0000 (16:12 +0800)]
MAINTAINERS: ARM MEDIATEK: update file entries

This patch adds new file entries for MediaTek SoCs

Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
5 years agoarm: MediaTek: add ethernet support for MT7629 boards
Weijie Gao [Thu, 20 Dec 2018 08:12:57 +0000 (16:12 +0800)]
arm: MediaTek: add ethernet support for MT7629 boards

Enable ethernet related configs to mt7629_rfb_defconfig.
Add default IP addresses.
Enable noncached memory region required by ethernet driver.

Signed-off-by: Mark Lee <Mark-MC.Lee@mediatek.com>
5 years agoarm: MediaTek: add ethernet support for MT7623 boards
Weijie Gao [Thu, 20 Dec 2018 08:12:56 +0000 (16:12 +0800)]
arm: MediaTek: add ethernet support for MT7623 boards

Enable ethernet related configs to mt7623n_bpir2_defconfig.
Add default IP addresses.
Enable noncached memory region required by ethernet driver.

Signed-off-by: Mark Lee <Mark-MC.Lee@mediatek.com>
5 years agoarm: dts: add ethernet related node for MT7629 SoC
Weijie Gao [Thu, 20 Dec 2018 08:12:55 +0000 (16:12 +0800)]
arm: dts: add ethernet related node for MT7629 SoC

This patch adds ethernet gmac node for MT7629 with internal gigabit phy.

Signed-off-by: Mark Lee <Mark-MC.Lee@mediatek.com>
5 years agoarm: dts: add ethernet related node for MT7623 SoC
Weijie Gao [Thu, 20 Dec 2018 08:12:54 +0000 (16:12 +0800)]
arm: dts: add ethernet related node for MT7623 SoC

This patch adds ethernet gmac node for MT7623 with MT7530 gigabit switch.

Signed-off-by: Mark Lee <Mark-MC.Lee@mediatek.com>
5 years agoethernet: MediaTek: add ethernet driver for MediaTek ARM-based SoCs
Weijie Gao [Thu, 20 Dec 2018 08:12:53 +0000 (16:12 +0800)]
ethernet: MediaTek: add ethernet driver for MediaTek ARM-based SoCs

This patch adds ethernet support for Mediatek ARM-based SoCs, including
a minimum setup of the integrated switch.

Cc: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Mark Lee <Mark-MC.Lee@mediatek.com>
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-By: "Frank Wunderlich" <frank-w@public-files.de>
5 years agoclk: MediaTek: bind ethsys reset controller
Weijie Gao [Thu, 20 Dec 2018 08:12:52 +0000 (16:12 +0800)]
clk: MediaTek: bind ethsys reset controller

The ethsys contains not only the clock gating controller, but also the
reset controller for the whole ethernet subsystem and its components.

This patch adds binding of the reset controller so that the ethernet node
can have references on it.

Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
5 years agoreset: MedaiTek: add reset controller driver for MediaTek SoCs
Weijie Gao [Thu, 20 Dec 2018 08:12:51 +0000 (16:12 +0800)]
reset: MedaiTek: add reset controller driver for MediaTek SoCs

This patch adds reset controller driver for MediaTek SoCs.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
5 years agopoplar_defconfig: enable fastboot support
Shawn Guo [Tue, 18 Dec 2018 09:52:07 +0000 (17:52 +0800)]
poplar_defconfig: enable fastboot support

It enables fastboot support on Poplar board by using DWC2 OTG gadget
driver.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
5 years agopoplar: add DWC2 OTG gadget support
Shawn Guo [Tue, 18 Dec 2018 09:52:06 +0000 (17:52 +0800)]
poplar: add DWC2 OTG gadget support

It enables DWC2 OTG gadget driver support for Poplar board.  As
usb2_phy_init() is being always called from board_init(), we can save
the call from board_usb_init().

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
5 years agoimage: fix compiling without CMD_FDT
Simon Goldschmidt [Mon, 17 Dec 2018 19:14:42 +0000 (20:14 +0100)]
image: fix compiling without CMD_FDT

Booting an image currently sets the environment variable "fdtaddr"
by calling into 'cmd/fdt.c'. As a result, linking U-Boot fails if
CMD_FDT is not enabled.

Fix this by adding 'if (CONFIG_IS_ENABLED(CMD_FDT))' to the two
places where 'set_working_fdt_addr()' is called.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoam57xx_evm_defconfig: Enable YMODEM support
Yan Liu [Mon, 17 Dec 2018 14:27:59 +0000 (09:27 -0500)]
am57xx_evm_defconfig: Enable YMODEM support

Enable CONFIG_SPL_YMODEM_SUPPORT to support UART boot

Signed-off-by: Yan Liu <yan-liu@ti.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agotest: bootcount: add bootcount-uclass test
Philipp Tomsich [Fri, 14 Dec 2018 20:14:29 +0000 (21:14 +0100)]
test: bootcount: add bootcount-uclass test

Add a test for the bootcount uclass, which uses the RTC bootcount backend
(i.e. drivers/bootcount/rtc.c is implictly also tested).

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoPrepare v2019.01 v2019.01
Tom Rini [Mon, 14 Jan 2019 22:02:36 +0000 (17:02 -0500)]
Prepare v2019.01

Signed-off-by: Tom Rini <trini@konsulko.com>
5 years agommc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()
Fabio Estevam [Mon, 19 Nov 2018 12:31:53 +0000 (10:31 -0200)]
mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

The following hang is observed on a Hummingboard 2 MicroSOM
i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)
Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1
Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial
---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

while (!(esdhc_read32(&regs->irqstat) & flags))
;

Instead of looping forever, provide an exit path so that a timeout
error can be propagated in the case irqstat does not report
any interrupts, which may happen when no eMMC is populated on
board.

Reported-by: Ricardo Salveti <rsalveti@rsalveti.net>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
Tested-by: Ricardo Salveti <rsalveti@rsalveti.net>
5 years agoimx8mq_evk_defconfig: Move file system options to Kconfig
Fabio Estevam [Sat, 29 Dec 2018 12:02:24 +0000 (10:02 -0200)]
imx8mq_evk_defconfig: Move file system options to Kconfig

Chris Spencer reports that when enabling ext4 read support without
also enabling write support the following error is seen:

fs/fs.c:198:12: error: 'ext4_write_file' undeclared here (not in a
function); did you mean 'ext4_read_file'?
   .write = ext4_write_file,
            ^~~~~~~~~~~~~~~
Fix this problem by moving these options to Kconfig.

Reported-by: Chris Spencer <spencercw@gmail.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
5 years agopoplar: save environment at a different offset
Alex Elder [Tue, 18 Dec 2018 03:08:18 +0000 (11:08 +0800)]
poplar: save environment at a different offset

Change CONFIG_ENV_OFFSET for Poplar to be 1MB further into the eMMC
than before.  This puts it immediately prior to the space we are
reserving offset 0x200000-0x400000 for UEFI to save its persistent
data.  Define CONFIG_ENV_SIZE as a product of env_mmc_nblks and the
sector size, like CONFIG_ENV_OFFSET is.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
5 years agoMerge branch 'master' of git://git.denx.de/u-boot-socfpga
Tom Rini [Fri, 11 Jan 2019 15:47:53 +0000 (10:47 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-socfpga

5 years agoMerge branch 'master' of git://git.denx.de/u-boot-usb
Tom Rini [Fri, 11 Jan 2019 15:47:41 +0000 (10:47 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-usb

5 years agoRevert "fs: fat: assign rootdir sector when accessing root directory"
Tom Rini [Fri, 11 Jan 2019 14:54:44 +0000 (09:54 -0500)]
Revert "fs: fat: assign rootdir sector when accessing root directory"

This particular commit is causing a regression on stih410-b2260 and
other platforms when reading from FAT16.  Noting that I had rebased the
original fix from Thomas onto then-current master, there is also
question from Akashi-san if the change is still needed after other FAT
fixes that have gone in.

This reverts commit a68b0e11ea774492713a65d9fd5bb525fcaefff3.

Reported-by: Patrice Chotard <patrice.chotard@st.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Thomas RIENOESSL <thomas.rienoessl@bachmann.info>
Signed-off-by: Tom Rini <trini@konsulko.com>
5 years agoRevert "dm: pinctrl: Prevent (re-)configuring pins when already done before relocation"
Lukasz Majewski [Wed, 9 Jan 2019 22:05:02 +0000 (23:05 +0100)]
Revert "dm: pinctrl: Prevent (re-)configuring pins when already done before relocation"

This reverts commit a7f4b4b344396590845e6552c82829ef68ef9f89.

As reported by Alex Kiernan the above optimization introduces a
regression in the below use case where:

1. Device has defined 'u-boot,dm-spl' property (@ eMMC DTS node)

2. The device downloads its MLO/SPL via UART (not eMMC - the eMMC pinmux
pins are NOT probed/configured in MLO/SPL).

3. The loaded via UART MLO/SPL wants to load Linux from eMMC. In this case
the DM core and pinctrl uclass checks 'u-boot,dm-spl' and don't
configure pins (as it thinks that those were initialized in MLO/SPL).

As we are very close to release - please revert this commit.

Reported-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoARM: dts: socfpga: Add missing SDMMC reset
Tien Fong Chee [Sun, 30 Dec 2018 08:13:45 +0000 (16:13 +0800)]
ARM: dts: socfpga: Add missing SDMMC reset

The SDMMC reset is missing from DT, so the reset manager cannot unreset
the SDMMC. Add the missing DT reset entry.

Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
5 years agodm: usb: gadget: Fix boot breakage on sunxi platforms
Jean-Jacques Hiblot [Fri, 21 Dec 2018 08:50:21 +0000 (09:50 +0100)]
dm: usb: gadget: Fix boot breakage on sunxi platforms

Fixes commit 013116243950 ("dm: usb: create a new UCLASS ID for USB gadget
devices")

The UCLASS_DRIVER for id UCLASS_USB_GADGET_GENERIC needs to be declared
even for platforms that do not enable DM_USB_GADGET. Otherwise the driver
for their usb peripheral controller fails to bind.

Reported-by: Priit Laes <plaes@plaes.org>
Reported-by: Jagan Teki <jagan@amarulasolutions.com>
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Tested-by: Priit Laes <plaes@plaes.org>
Acked-by: Jagan Teki <jagan@openedev.com>
5 years agousb: Make compiling gadget support optional
Jean-Jacques Hiblot [Thu, 10 Jan 2019 14:44:14 +0000 (15:44 +0100)]
usb: Make compiling gadget support optional

There is no need to compile and include this code if it is not used.
CONFIG_USB_GADGET can be used for the purpose.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
5 years agoKconfig: rename CONFIG_SPL_USB_GADGET_SUPPORT as CONFIG_SPL_USB_GADGET
Jean-Jacques Hiblot [Thu, 10 Jan 2019 14:44:13 +0000 (15:44 +0100)]
Kconfig: rename CONFIG_SPL_USB_GADGET_SUPPORT as CONFIG_SPL_USB_GADGET

The SPL option for USB gadget should be named after the option for u-boot
(CONFIG_USB_GADGET)

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
5 years agoARM: dts: define USB aliases for all omap5 platforms
Jean-Jacques Hiblot [Sat, 15 Dec 2018 16:43:28 +0000 (17:43 +0100)]
ARM: dts: define USB aliases for all omap5 platforms

This allows us to properly map the USB controller indexes

Tested on dra76 evm, am572 evm

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
5 years agodm: usb: udc: Use SEQ_ALIAS to index the USB gadget ports
Jean-Jacques Hiblot [Sat, 15 Dec 2018 16:43:27 +0000 (17:43 +0100)]
dm: usb: udc: Use SEQ_ALIAS to index the USB gadget ports

dfu, fastbot and other usb gadget commands take the USB port index as a
parameter. Currently this index is assigned in the order of the driver
bindings.
Changing this behavior using the SEQ_ALIAS feature. This option assign to
the device a SEQ number based on its alias (if it exists)

To use it we must set the DM_UC_FLAG_SEQ_ALIAS flag and follow the existing
naming convention: use "usb" for the name of the gadget UCLASS_DRIVER
(same as for the UCLASS_USB).

If no alias is provided, then the index falls back to the order in which
the bindings took place.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reported-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Tested-by: Sam Protsenko <semen.protsenko@linaro.org>
5 years agoMerge branch 'master' of git://git.denx.de/u-boot-sunxi
Tom Rini [Thu, 10 Jan 2019 14:28:28 +0000 (09:28 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-sunxi

5 years agoMerge tag 'u-boot-imx-20190110' of git://git.denx.de/u-boot-imx
Tom Rini [Thu, 10 Jan 2019 14:28:16 +0000 (09:28 -0500)]
Merge tag 'u-boot-imx-20190110' of git://git.denx.de/u-boot-imx

Fixes for 2019.01

5 years agommc: sunxi: Fix mmc clocks for DM_MMC
Jagan Teki [Wed, 9 Jan 2019 11:28:39 +0000 (16:58 +0530)]
mmc: sunxi: Fix mmc clocks for DM_MMC

Existing clock configure code has been followed based on the
legacy MMC dt node definitions and it cannot work with recent
dts(i) sync from Linux.

So, add clock configure code for Allwinner platforms which support
DM_MMC and eventually this will drop once CLK support is in Mainline.

Fixes: 3c92cca3cda0 ("ARM: dts: sun4i: Update A10 dts(i) files from Linux-v4.18-rc3")
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Priit Laes <plaes@plaes.org> # Gemei G9 A10 Tablet
Tested-by: Marek Kraus <gamelasterv2@gmail.com> # A10-OLinuXino-Lime
5 years agoMerge branch '2019-01-08-master-imports'
Tom Rini [Wed, 9 Jan 2019 21:02:29 +0000 (16:02 -0500)]
Merge branch '2019-01-08-master-imports'

- stm32f7 GPIO fixes
- SATA env fixes
- More DM migration deadline warnings
- Regression fix for non-DM MMC drivers
- dma_alloc_coherent size fix on ARM.

5 years agoimx8m: clock: Fix oscillator values
Fabio Estevam [Fri, 28 Dec 2018 18:43:01 +0000 (16:43 -0200)]
imx8m: clock: Fix oscillator values

OSC_27M_CLK should return 27MHz and OSC_32K_CLK should return
32768Hz to reflect the reality.

This also keeps the values in sync with the Linux clock tree.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
5 years agotools: imx8image: set dcd_skip to true
Peng Fan [Sat, 15 Dec 2018 12:19:58 +0000 (12:19 +0000)]
tools: imx8image: set dcd_skip to true

To B0[+] chips, dcd_skip needs to be true. For A0 chip, it needs
to be false, however A0 chip is no longer being supported anymore.
Considering we are moving code from imx-mkimage to uboot mkimage,
to make sure we not introduce some surprise, we still keep dcd_skip
code there.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
5 years agoimx8: cpu: correct info
Peng Fan [Sat, 15 Dec 2018 12:19:55 +0000 (12:19 +0000)]
imx8: cpu: correct info

The CPU banner printed is as following:
CPU:   CPU:   Freescale i.MX8QXP RevB A35 at 147228 MHz

1. Drop the CPU:
2. Change vendor from Freescale to NXP

Signed-off-by: Peng Fan <peng.fan@nxp.com>
5 years agomisc: imx8: scu: use platdata instead of priv data
Peng Fan [Sat, 15 Dec 2018 12:19:52 +0000 (12:19 +0000)]
misc: imx8: scu: use platdata instead of priv data

priv data has not been allocated when doing bind, so it is
wrong to use dev_get_priv in bind call back.

Let's switch to use platdata in the driver to fix the issue.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agomisc: imx: scu: avoid write null pointer
Peng Fan [Sat, 15 Dec 2018 12:19:49 +0000 (12:19 +0000)]
misc: imx: scu: avoid write null pointer

When boot_dev is true, fill boot device. However the original logic is when
boot_dev is false, fill boot device, this will trigger data abort.

Also fix sc_misc_get_control when using pointer val.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
5 years agoclk: imx8: fix build warning
Peng Fan [Sat, 15 Dec 2018 12:19:46 +0000 (12:19 +0000)]
clk: imx8: fix build warning

When build clk driver in spl, met the warning:
"
drivers/clk/imx/clk-imx8.c:21:25: warning: ‘imx8_clk_names’ defined but not used [-Wunused-variable]
 static struct imx8_clks imx8_clk_names[] = {
                         ^~~~~~~~~~~~~~
"

Fix with wrapping the array with CONFIG_CMD_CLK.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
5 years agoMAINTAINERS: update NXP i.MX mail list address
Peng Fan [Sat, 15 Dec 2018 12:07:02 +0000 (12:07 +0000)]
MAINTAINERS: update NXP i.MX mail list address

Update NXP i.MX mail list address

Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
5 years agoboard: tbs2910: Remove FIT support in defconfig to reduce u-boot size
Soeren Moch [Sat, 5 Jan 2019 08:31:18 +0000 (09:31 +0100)]
board: tbs2910: Remove FIT support in defconfig to reduce u-boot size

The current defconfig build generates a u-boot.imx file that is too large
for the available space on a eMMC/SD card. Installing this file overwrites
the u-boot environment. So disable the unused FIT support to reduce the
size of the u-boot binary.

Signed-off-by: Soeren Moch <smoch@web.de>
5 years agoboard: tbs2910: Add u-boot.imx size limit check
Soeren Moch [Sat, 5 Jan 2019 08:31:17 +0000 (09:31 +0100)]
board: tbs2910: Add u-boot.imx size limit check

Check the size of the generated u-boot.imx file. Report an error
if it would be too big and overwrite the u-boot environment.

Signed-off-by: Soeren Moch <smoch@web.de>
5 years agoimx: Add Makefile dependency for mkimage_fit_atf.sh
Chris Spencer [Thu, 20 Dec 2018 09:25:24 +0000 (09:25 +0000)]
imx: Add Makefile dependency for mkimage_fit_atf.sh

The mkimage_fit_atf.sh SPL FIT generator script requires
u-boot-nodtb.bin, but this was not enforced by the Makefile. This could
cause the generator script to be executed before u-boot-nodtb.bin has
been created.

Signed-off-by: Chris Spencer <christopher.spencer@sea.co.uk>
Cc: NXP i.MX U-Boot Team <uboot-imx@nxp.com>
5 years agoarm: mx5: Enable WDT and bootcounter on M53Menlo
Marek Vasut [Fri, 7 Dec 2018 01:36:38 +0000 (02:36 +0100)]
arm: mx5: Enable WDT and bootcounter on M53Menlo

Enable watchdog and bootcounter support on the M53Menlo board.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
5 years agoARM: vf610: ddrmc: do not write CR79 by default
Stefan Agner [Tue, 4 Dec 2018 10:10:21 +0000 (11:10 +0100)]
ARM: vf610: ddrmc: do not write CR79 by default

The current value CTLUPD_AREF(0) is the reset value of the register,
so there is no need to write a value. If needed, the register can be
written using board specific CR settings.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agoARM: vf610: ddrmc: fix initialization completion detection
Stefan Agner [Tue, 4 Dec 2018 10:10:20 +0000 (11:10 +0100)]
ARM: vf610: ddrmc: fix initialization completion detection

The CR80 register has multiple interrupt bits, the code is supposed
to check bit 8 but instead uses a logical and. In most cases this
probably did not affect real operations since at that stage typically
none of the other bits are set.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
5 years agoARM: vf610: ddrmc: fix CR138 preprocessor define
Stefan Agner [Tue, 4 Dec 2018 10:10:19 +0000 (11:10 +0100)]
ARM: vf610: ddrmc: fix CR138 preprocessor define

According to the data sheet bits 10-8 are PHYDRAM_CK_EN. Fix mask
to allow setting PHYDRAM_CK_EN correctly.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agotoradex: colibri_vf: fix memory initialization
Stefan Agner [Tue, 4 Dec 2018 10:10:18 +0000 (11:10 +0100)]
toradex: colibri_vf: fix memory initialization

Commit 3f353ceccbbb ("vf610: refactor DDRMC code") changed on-die
termination (ODT) values from 120 Ohm to 60 Ohm and enabled a static
read/write leveling which has not been tested with this board. This
commit reverts both changes and makes sure that memory gets
initialized as it has been done before the mentioned commit.

Fixes: 3f353ceccbbb ("vf610: refactor DDRMC code")
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
5 years agoARM: vf610: ddrmc: program Dummy DDRBYTE1/2
Stefan Agner [Fri, 14 Dec 2018 14:26:00 +0000 (15:26 +0100)]
ARM: vf610: ddrmc: program Dummy DDRBYTE1/2

The Vybrid reference manual VFXXXRM Rev. 0 10/2016 states in chapter
5.2.6.1 DUMMY PADS (DDR/QuadSPI) that those pads need to be programed
for correct operation of DDR. Assume the default DDR pin configuration
which seems to work well on a Colibri VF50.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
5 years agoboard: toradex: colibri_vf: unset NFS and LOADS/B
Stefan Agner [Thu, 6 Dec 2018 13:55:03 +0000 (14:55 +0100)]
board: toradex: colibri_vf: unset NFS and LOADS/B

Safe some space by not selecting CMD_NFS and CMD_LOADS/B.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: xhci-mem: Fix scratchpad array issue
Ye Li [Mon, 7 Jan 2019 02:45:46 +0000 (02:45 +0000)]
usb: xhci-mem: Fix scratchpad array issue

After updating the value of dev_context_ptrs[0], we should flush this
from cache to memory. Otherwise the xhci controller won't use it.

Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Marek Vasut <marex@denx.de>
5 years agogpio: stm32f7: Fix SPL code size
Patrice Chotard [Fri, 4 Jan 2019 09:55:06 +0000 (10:55 +0100)]
gpio: stm32f7: Fix SPL code size

In order to keep SPL code size below the 32Kb limit,
put under CONFIG_SPL_BUILD flag all unused code in SPL.
This is needed for stm32f7xx board which are using SPL.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
5 years agogpio: stm32f7: Fix gpio bank hole management
Patrice Chotard [Fri, 4 Jan 2019 09:55:05 +0000 (10:55 +0100)]
gpio: stm32f7: Fix gpio bank hole management

In case "gpio-ranges" property is not present in device tree,
use default value for gpio_count and gpio_range.
This fixes an issue on stm32 F7 and H7 boards where "pinmux status -a"
command didn't return any pin status due to the fact that both stm32 F7
and H7 board DT doesn't use the gpio-ranges property.

Fixes: dbf928dd2634a6("gpio: stm32f7: Add gpio bank holes management")

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
5 years agoenv: sata: Fix saveenv issue
Ye Li [Mon, 7 Jan 2019 09:22:35 +0000 (09:22 +0000)]
env: sata: Fix saveenv issue

Wrong env buffer was passed into sata write function, cause the saveenv
not work.

Signed-off-by: Ye Li <ye.li@nxp.com>
5 years agoarm: ti: boot: Remove legacy Android partitions
Sam Protsenko [Fri, 4 Jan 2019 20:09:08 +0000 (22:09 +0200)]
arm: ti: boot: Remove legacy Android partitions

Remove unused Android partitions:
  - efs, crypto, cache: we don't use it anymore (images are not built
    in AOSP
  - ipu1, ipu2: IPU firmware is now a part of vendor image and doesn't
    reside as a separate partition

While at it, rename "reserved" partition to "uboot-env", as it's
actually stores U-Boot environment.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Praneeth Bajjuri <praneeth@ti.com>
5 years agoenv: sata: Add missed env location for SATA boot
Ye Li [Fri, 4 Jan 2019 09:34:24 +0000 (09:34 +0000)]
env: sata: Add missed env location for SATA boot

The env location label ENVL_ESATA is missed in location tables, so
when we configure the ENV in SATA, u-boot fails to get correct env
location and cause boot hang in board_f.

Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
5 years agoarm: Round the dma_alloc_coherent memory size to cache line aligned
Ye Li [Fri, 4 Jan 2019 09:24:14 +0000 (09:24 +0000)]
arm: Round the dma_alloc_coherent memory size to cache line aligned

When running usb dwc3 gadget driver, we meet random USB enumeration failure in fastboot.
The root cause is a cache coherence issue. When it happens, the ctrl_req in
gadget driver is allocated at 0xfe932f40, and the usb_composite_dev (cdev)
is allocated at 0xfe932f60. So after we submit the setup request (cache flushed) to USB
controller, any accessing to usb_composite_dev variable will cause the cache line refill, then
when setup transfer is completed, reading the setup data in ctrl_req will gets old value from
cache not from memory.

The ctrl_req is allocated by API dma_alloc_coherent, but u-boot don't have cohernet memory.
so it still needs cache maintain operations before/after HW accessing. Since the cache flush or
invalidate bases on cache line, so when the allocated memory size is not cache line aligned,
potentially it may meet such issue.

This patch modifies the dma_alloc_coherent API to round the size to cache line aligned.

Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
5 years agodoc: README.commands: fix type
Heinrich Schuchardt [Sun, 30 Dec 2018 12:00:51 +0000 (13:00 +0100)]
doc: README.commands: fix type

%s/commmand/command/

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agodm: video: lcd: MIGRATION: Add migration plan for video
Simon Glass [Fri, 28 Dec 2018 21:03:08 +0000 (14:03 -0700)]
dm: video: lcd: MIGRATION: Add migration plan for video

Add a migration plan for video which is an important subsystem in U-Boot.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agodm: pci: MIGRATION: Add migration plan for PCI
Simon Glass [Fri, 28 Dec 2018 21:03:07 +0000 (14:03 -0700)]
dm: pci: MIGRATION: Add migration plan for PCI

Add a migration plan for PCI which is an important subsystem in U-Boot.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agoigep003x: MAINTAINER: Remove myself as maintainer and add Javier
Enric Balletbo i Serra [Fri, 28 Dec 2018 10:55:47 +0000 (11:55 +0100)]
igep003x: MAINTAINER: Remove myself as maintainer and add Javier

I can't continue maintaining the board because I don't have access to the
hardware anymore, so remove myself from the entry and add Javier who has
volunteered to help and maintain the board.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Javier Martínez Canillas <javier@dowhile0.org>
5 years agoAdd a github template telling people to not use pull requests there
Tom Rini [Thu, 27 Dec 2018 14:22:55 +0000 (09:22 -0500)]
Add a github template telling people to not use pull requests there

On our mirror account on github we have gotten a handful of pull
requests.  At this time github does not allow you to disable pull
requests on a project.  The generally suggested work-around is to add a
pull request template that tells people to not use that workflow.  Add
one here that points to the wiki page on submitted patches.

Signed-off-by: Tom Rini <trini@konsulko.com>
5 years agommc: bring back partition init for non-DM MMC drivers
Andre Przywara [Mon, 17 Dec 2018 10:05:45 +0000 (10:05 +0000)]
mmc: bring back partition init for non-DM MMC drivers

Commit d0851c893706 ("blk: Call part_init() in the post_probe() method")
removed the call to part_init() in mmc.c, as this is done by the DM_MMC
framework.
However Allwinner is (still) relying on a non-DM MMC driver, so we are
now missing the implicit partition init, leading to failing MMC accesses
due to the missing partition information.

Bring the call back just for non-DM MMC driver to fix this regression.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Soeren Moch <smoch@web.de>
5 years agosunxi: drop default SPL_LIBDISK_SUPPORT enablement
Andre Przywara [Mon, 17 Dec 2018 10:05:44 +0000 (10:05 +0000)]
sunxi: drop default SPL_LIBDISK_SUPPORT enablement

There is no code for using partition labels in the Allwinner SPL port.
Even so the name is slightly misleading, CONFIG_SPL_LIBDISK_SUPPORT was
meant to guard partition code for the SPL.

Remove the "imply" line in the Kconfig to make this obvious and avoid
unneeded code inclusions, helping to keep the H6 SPL code small.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoMakefile: Correct logic for DM_SCSI + unconverted drivers check
Tom Rini [Tue, 8 Jan 2019 03:57:56 +0000 (22:57 -0500)]
Makefile: Correct logic for DM_SCSI + unconverted drivers check

When checking for boards that are enabling a SATA driver that isn't
converted to DM yet we need to be sure to not also trip over boards that
do set CONFIG_DM_SCSI by itself, as that is not a bug.

Reported-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Fixes: ea9d7c17fc4c ("dm: MIGRATION: Add migration plan for CONFIG_SATA")
Signed-off-by: Tom Rini <trini@konsulko.com>