10 Feb, 2017
9 commits
-
The kernel partition in QSPI is 0x440000 large, not 0x400000. Fix this
in the environment, otherwise the kernel will fail to boot if it occupies
more space.Signed-off-by: Mike Looijmans
Signed-off-by: Michal Simek -
Allow sending firmware to RAM. Without this, the DFU support was not
of much use.Signed-off-by: Mike Looijmans
Signed-off-by: Michal Simek -
On the miami board, ethernet is accessed via logic. To use it, one
would have to program logic first and then set up the rgmii conversion
block as well. Not likely to ever be used, so disable network support
by default to save some space.Signed-off-by: Mike Looijmans
Signed-off-by: Michal Simek -
Keep all defconfig sorted to ensure the smallest diff.
Signed-off-by: Michal Simek
-
It was incorrect to always include "asm/arch-omap3/mux.h" constantly.
This introduced warnings on non-omap3 where certain values will conflict
between the various families. Conditionally guard the inclusion in
order to correct the problem.Fixes: 6aca17c9b7e8 ("drivers: mmc: omap_hsmmc: Fix IO Buffer on OMAP36xx")
Signed-off-by: Tom Rini -
To save more space, switch to simple malloc here.
Signed-off-by: Tom Rini
09 Feb, 2017
31 commits
-
The qemu-x86* targets do not want to enable this.
Signed-off-by: Tom Rini
-
MMC devices accessed exclusively via the driver model were not
being initialized before being exposed as block devices, causing
issues in scenarios where the MMC device is first accessed via the
uclass block interface.Signed-off-by: Fiach Antaw
-
On the OMAP36xx/37xx the CONTROL_WKUP_CTRL register has
a field (bit 6) named GPIO_IO_PWRDNZ. If 0, the IO buffers which
are related to GPIO_126, 127 and 129 are disabled. Some boards may
need this for MMC. After the PBIAS is configured, this bit should
be set high to enable these GPIO pins.Signed-off-by: Adam Ford
Reviewed-by: Tom Rini -
ftsdc021_sdhci.c is dead file.
There is no reason to maintain this host controller.
Removes the entire ftsdc021_sdhci.c.Signed-off-by: Jaehoon Chung
Reviewed-by: Tom Rini -
If there are alias nodes as "mmc", use the devnum as alias index
number.
This patch is for fixing a problem of Exynos4 series.
Problem is the below thing.Current legacy mode:
EXYNOS DWMMC: 0, SAMSUNG SDHCI: 1After using DM:
SAMSUNG SDHCI: 0, EXYNOS DWMMC: 1Dev index is swapped.
Then u-boot can't find the kernel image..because it is already set to 0 as mmcdev.
If change from legacy to DM, also needs to touch all exynos4 config file.
For using simply, just supporting the fixed devnum with alias node is better than it.Usage:
alaise {
....
mmc0 = &sdhci2; /* eMMC */
mmc1 = &sdhci1; /* SD */
...
}Signed-off-by: Jaehoon Chung
Reviewed-by: Simon Glass -
To use driver-model adds the pmic node for max8997.
This is used as kernel device-tree in Linux.Signed-off-by: Jaehoon Chung
Reviewed-by: Simon Glass -
Add the max8997 controller for Driver model.
Exynos4210 is using max8997 pmic controller.
(pmic_max8997.c should be deprecated.)Signed-off-by: Jaehoon Chung
Reviewed-by: Simon Glass -
Signed-off-by: John Haechten
Acked-by: Joe Hershberger -
Signed-off-by: Robert P. J. Day
-
Albeit it's a typo, neither CONGIG_CMD_STORAGE nor CONFIG_CMD_STORAGE
are used anywhere, so remove the define from the board configs.Signed-off-by: Lars Poeschel
Reviewed-by: Tom Rini -
CONFIG_CMD_ZIP is not defined by any board. I am moving
CONFIG_CMD_UNZIP to defconfig files except UniPhier SoC family.I am the maintainer of UniPhier platform, so I know "select CMD_UNZIP"
is better for this platform.Signed-off-by: Masahiro Yamada
Acked-by: Michal Simek
Acked-by: Stefan Roese
Acked-by: Ryan Harkin
Tested-by: Ryan Harkin -
There are two typos in the comment "invalide i-cache is enabled".
We can fix it by
invalide -> invalidate
is -> ifOr, if we want to match the comment to the code, we can say
"skip invalidating i-cache if disabled".Signed-off-by: Masahiro Yamada
Reviewed-by: Simon Glass -
CONFIG_SPL_STACK_SIZE is not a config option, so rename it.
Signed-off-by: Phil Edworthy
-
The array indices used currently are dispalaced by 1 for
SMPS6 through SMPS10 in the respective places of voltage and ctrl
arrays hence fix the same as to assign the right voltage and ctrl
registers.Signed-off-by: Keerthy
-
Both CONFIG_PWM_TEGRA and CONFIG_PWM_EXYNOS depend on CONFIG_DM_PWM,
i.e. they are already guarded by Kconfig correctly. Remove unneeded
ifdef CONFIG_DM_PWM ... endif.While we are here, let's tidy up alignment and sort the lines
alphabetically in Makefile.Signed-off-by: Masahiro Yamada
Reviewed-by: Simon Glass -
The default values for the configuration defines CONFIG_ENV_SPI_xxx are
arbitrary values. It makes more sense to set them to the values used by
the sf command.Signed-off-by: Jean-Jacques Hiblot
Reviewed-by: Tom Rini -
The sequence erroneously launched the DDR controller
initialization before the pad muxing was done, causing
DRAM size computation to hang.Configuring the pads first then launching DDR controller
initialization prevents the DRAM hanging.Signed-off-by: Albert ARIBAUD (3ADEV)
-
cpsw driver tries to get macid for am43xx SoCs using the compatible
ti,am4372. But not all variants of am43x uses this complatible like
epos evm uses ti,am438x. So use a generic compatible ti,am43 to get
macid for all am43 based platforms.Tested-by: Aparna Balasubramanian
Signed-off-by: Lokesh Vutla
Reviewed-by: Joe Hershberger
Reviewed-by: Tom Rini -
Now when CONFIG_CMD_IMLS_NAND is enabled the u-boot build will fail,
because nand_read_skip_bad() function has been changed to accept more
parameters, hence fix it.CC cmd/bootm.o
cmd/bootm.c: In function 'nand_imls_legacyimage':
cmd/bootm.c:390:8: error: too few arguments to function 'nand_read_skip_bad'
ret = nand_read_skip_bad(mtd, off, &len, imgdata);
^
In file included from cmd/bootm.c:18:0:
include/nand.h:101:5: note: declared here
int nand_read_skip_bad(struct mtd_info *mtd, loff_t offset, size_t *length,
^
LD drivers/block/built-in.oSigned-off-by: Grygorii Strashko
Reviewed-by: Tom Rini
Reviewed-by: Simon Glass -
Pick commit 66b47b4a9dad0 checkpatch: look for common misspellings
from the Linux kernel for spelling check from Kees CookIn addition pulled in additional changes
commit ebfd7d6237531 checkpatch: add optional --codespell dictionary to find more typos
from the Linux kernel for codespell from Joe Perchescommit f1a63678554f8 checkpatch: remove local from codespell path
from the Linux kernel for dictionary path from Maxim UvarovSigned-off-by: Dan Murphy
Reviewed-by: Tom Rini -
Now that we have a Kconfig for the API, convert the two boards that
are using this to Kconfig and remove CONFIG_API from the whitelist.Signed-off-by: Emmanuel Vadot
Reviewed-by: Tom Rini
Reviewed-by: Simon Glass -
Add kconfig file to enable API support
Signed-off-by: Emmanuel Vadot
Reviewed-by: Tom Rini
Reviewed-by: Simon Glass -
Use the kbuild style log.
Prior to this commit:
./scripts/check-config.sh u-boot.cfg \
./scripts/config_whitelist.txt . 1>&2With this commit:
CFGCHK u-boot.cfg
Signed-off-by: Masahiro Yamada
Reviewed-by: Simon Glass -
As reported in [1], rename the k2* dts files to keystone-* files
this will force consistency throughout.Script for the same (and hand modified for Makefile and config
files):
for i in arch/arm/dts/k2*
do
b=`basename $i`;
git mv $i arch/arm/dts/keystone-$b;
sed -i -e "s/$b/keystone-$b/g" arch/arm/dts/*[si]
doneThis is similar to linux kernel commit 5edafc29829bc ("ARM: dts: k2*: Rename
the k2* files to keystone-k2* files")[1] http://marc.info/?l=linux-arm-kernel&m=145637407804754&w=2
Signed-off-by: Lokesh Vutla
Reviewed-by: Tom Rini -
Fix H-PLL and M-PLL rate calculation in ast2500 clock driver.
Without this fix, valid setting can lead to division by zero
when requesting the rate of H-PLL or M-PLL clocks.Signed-off-by: Maxim Sloyko
Reviewed-by: Simon Glass -
On some cases the first 440 bytes of MBR are used to keep an additional
information for ROM boot loader. 'gpt write' command doesn't preserve
that area and makes boot code gone.Preserve boot code area when run 'gpt write' command.
Signed-off-by: Vincent Tinelli
Signed-off-by: Brennan Ashton
Signed-off-by: Andy Shevchenko
Reviewed-by: Simon Glass -
The previous commit ("add a new command to show .config contents")
improves the basic infrastructure of "embed a compressed file into
the U-Boot image, and print it by a command". The same pattern for
the "license" command.This commit reworks the command to improve the following:
[1] Improve log style
Kbuild style log
GZIP cmd/license_data.gz
CHK cmd/license_data_gz.h
UPD cmd/license_data_gz.h
CHK cmd/license_data_size.h
UPD cmd/license_data_size.hinstead of the bare Make log:
cat ./Licenses/gpl-2.0.txt | gzip -9 -c | \
tools/bin2header license_gzip > ./include/license.h[2] Collect related code into the "cmd" directory
Prior to this commit, the license.h was created by tools/Makefile,
placed under the "include" directory, included from cmd/license.c,
and deleted by the top-level Makefile. It is not a good idea to
scatter related code.[3] Drop the fixed-malloc size LICENSE_MAX
Just allocate the minimum required size of buffer because we know
the size of the original gpl-2.0.txt.[4] Fix more issues
Terminate the buffer with zero to prevent puts() from over-running.
Add "static" to do_license.Signed-off-by: Masahiro Yamada
Reviewed-by: Simon Glass -
This feature is inspired by /proc/config.gz of Linux. In Linux,
if CONFIG_IKCONFIG is enabled, the ".config" file contents are
embedded in the kernel image. If CONFIG_IKCONFIG_PROC is also
enabled, the ".config" contents are exposed to /proc/config.gz.
Users can do "zcat /proc/config.gz" to check which config options
are enabled on the running kernel image.The idea is almost the same here; if CONFIG_CMD_CONFIG is enabled,
the ".config" contents are compressed and saved in the U-Boot image,
then printed by the new command "config".The usage is quite simple. Enable CONFIG_CMD_CONFIG, then run
> config
from the command line interface. The ".config" contents will be
printed on the console.This feature increases the U-Boot image size by about 4KB (this is
mostly due to the gzip-compressed .config file). By default, it is
enabled only for Sandbox because we do not care about the memory
footprint on it. Of course, this feature is architecture agnostic,
so you can enable it on any board if the image size increase is
acceptable for you.Signed-off-by: Masahiro Yamada
Reviewed-by: Simon Glass -
Import scripts/basic/bin2c.c of Linux.
In Linux Kernel, this file was moved to scripts/basic directory by
commit 8370edea81e3 ("bin2c: move bin2c in scripts/basic").In U-Boot, we do not need to follow that commit. Just put it in the
original directory "scripts".Signed-off-by: Masahiro Yamada
Reviewed-by: Simon Glass