Masahiro Yamada [Thu, 21 Apr 2016 05:43:11 +0000 (14:43 +0900)]
pinctrl: uniphier: fix NAND and SD pin-mux settings for PH1-LD11/LD20
I found many mistakes in the initial version.
Fixes:
8a3328c209d0 ("pinctrl: uniphier: support UniPhier PH1-LD20 pinctrl driver")
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Masahiro Yamada [Thu, 21 Apr 2016 05:43:10 +0000 (14:43 +0900)]
pinctrl: uniphier: rename function/array names
Make function/array names match the file names for consistency.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Tom Rini [Fri, 22 Apr 2016 11:25:18 +0000 (07:25 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-x86
Heiko Schocher [Thu, 21 Apr 2016 10:16:58 +0000 (12:16 +0200)]
ubifs: fix memory corruption in super.c
In list "super_blocks" ubifs collects allocated super_block
structs. U-Boot frees on unmount the allocated struct,
so the pointer stored in this list is free after the umount.
On a new ubifs mount, the new allocated super_block struct
get inserted into the super_blocks list ... which contains
now a freed pointer, and the list_add_tail() corrupts the
freed memory ...
2 solutions are possible:
- remove the super_block from the super_blocks list
on umount
- as U-Boot does not use the super_blocks list ...
remove it complete for U-Boot.
Both solutions should not introduce problems for porting
to newer linux version, so this patch removes the unused
super_blocks list, as it saves code size and execution
time.
Signed-off-by: Heiko Schocher <hs@denx.de>
Heiko Schocher [Tue, 2 Feb 2016 10:54:35 +0000 (11:54 +0100)]
mtd, ubi: set free_count to zero before walking through erase list
Set free_count to zero before walking through ai->erase list
in wl_init().
As U-Boot has no workqueue/threads, it immediately calls
erase_worker(), which increase for each erased block
free_count. Without this patch, free_count gets after
this initialized to zero in wl_init(), so the free_count
variable always has the maybe wrong value 0.
Detected this behaviour on the dxr2 board, where the
UBI fastmap gets not written when attaching/dettaching
on an empty NAND. It drops instead the error message:
could not find any anchor PEB
With this patch, fastmap gets written on dettach.
Signed-off-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Bin Meng [Wed, 13 Apr 2016 08:00:51 +0000 (01:00 -0700)]
x86: Correct typo of Miao Yan's email address
Miao Yan's email address is wrong in fw_cfg.c. Fix it.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Bin Meng [Wed, 13 Apr 2016 07:52:49 +0000 (00:52 -0700)]
x86: qemu: Drop our own ACPI implementation
Our own ACPI implementation (when CONFIG_QEMU_ACPI_TABLE is not set)
does not build anymore after x86 has been fully converted to DM PCI.
Instead of trying to fix the build errors, given we now have the ACPI
support via QEMU's fw_cfg interface, which is a more reliable way to
generate correct ACPI tables than by ourselves, hence drop our own
ACPI implementation.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Alexey Brodkin [Sat, 16 Apr 2016 12:28:30 +0000 (15:28 +0300)]
arc/cache: really do flush_dcache_all() even if IOC exists
flush_dcache_all() is used in the very end of U-Boot self relocation
to write back all copied and then patched code and data to their
new location in the very end of available memory space.
Since that has nothing to do with IO (i.e. no external DMA happens
here) IOC won't help here and we need to write back data cache contents
manually.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Tom Rini [Thu, 21 Apr 2016 13:37:33 +0000 (09:37 -0400)]
Prepare v2016.05-rc2
Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Wed, 20 Apr 2016 19:33:05 +0000 (15:33 -0400)]
configs: Update some Xilinx configs
As part of 3457bba these configs didn't get updated. Update them now.
Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Wed, 20 Apr 2016 14:59:32 +0000 (10:59 -0400)]
cmd/usb_mass_storage.c: Rework ums_init() ret logic slightly
Previously, ret could be used uninitialized if
blk_get_device_part_str() failed. Default to ret being set to -1 so
that we always return an err up if we have a problem and then invert the
logic on testing ums_count as when that is non-zero is the time we can
return 0.
Cc: John Tobias <john.tobias.ph@gmail.com>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Wed, 20 Apr 2016 13:31:58 +0000 (09:31 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-usb
Conflicts:
configs/bcm28155_ap_defconfig
configs/dra72_evm_defconfig
configs/dra74_evm_defconfig
configs/ma5d4evk_defconfig
Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Wed, 20 Apr 2016 13:23:42 +0000 (09:23 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-socfpga
Sam Protsenko [Wed, 13 Apr 2016 11:20:30 +0000 (14:20 +0300)]
usb: gadget Move: CONFIG_G_DNL_* to Kconfig
And also reformat defconfigs using "make savedefconfig" rule.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Sam Protsenko [Wed, 13 Apr 2016 11:20:29 +0000 (14:20 +0300)]
usb: dwc3: Move CONFIG_USB_DWC3_PHY_* to Kconfig
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Sam Protsenko [Wed, 13 Apr 2016 11:20:28 +0000 (14:20 +0300)]
usb: dwc3: Move CONFIG_USB_DWC3_OMAP to Kconfig
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Sam Protsenko [Wed, 13 Apr 2016 11:20:27 +0000 (14:20 +0300)]
usb: dwc3: Move CONFIG_USB_DWC3_GADGET/HOST to Kconfig
Description was borrowed from kernel dwc3 Kconfig.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Sam Protsenko [Wed, 13 Apr 2016 11:20:26 +0000 (14:20 +0300)]
usb: gadget: Move CONFIG_USB_GADGET_DOWNLOAD to Kconfig
While at it, remove obsolete CONFIG_USBDOWNLOAD_GADGET option from some
config headers. This is also probably fixes am335x_baltos board.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Sam Protsenko [Wed, 13 Apr 2016 11:20:25 +0000 (14:20 +0300)]
usb: gadget: Move CONFIG_USB_GADGET_DUALSPEED to Kconfig
Move CONFIG_USB_GADGET_DUALSPEED option to Kconfig and
make all UDC controllers select USB_GADGET_DUALSPEED:
- add next options to Kconfig selecting USB_GADGET_DUALSPEED:
- USB_GADGET_ATMEL_USBA
- USB_GADGET_DWC2_OTG
- USB_DWC3
- CI_UDC
- make USB_MUSB_GADGET select USB_GADGET_DUALSPEED
While at it, make some related fixes:
- remove DUALSPEED from configs that don't enable gadget support:
- kwb.h
- tseries.h
- add missing USB_GADGET option to next configs:
- novena_defconfig
- pcm051_rev*_defconfig
- xfi3_defconfig
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Sam Protsenko [Wed, 13 Apr 2016 11:20:24 +0000 (14:20 +0300)]
usb: gadget: Move CONFIG_USB_GADGET_VBUS_DRAW to Kconfig
The description was borrowed from kernel. Definitions were added to
defconfig files in a way that "make savedefconfig" generates exactly
the same file as used defconfig.
Boards using 0 mA as CONFIG_USB_GADGET_VBUS_DRAW value were moved to use
2 mA (as minimal allowed by Kconfig).
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
John Tobias [Wed, 13 Apr 2016 19:06:51 +0000 (12:06 -0700)]
usb: ums - expose selected partition/s
By applying this patch, it will give us some flexibility to expose
a selected partition/s.
e.g:
1. To expose several partitions
ums 0 mmc 0:1,0:6
2. To expose the all partitions
ums 0 mmc 0:0
3. To expose multiple partititions on several devices
ums 0 mmc 0:1,1:6
4. It support legacy format
ums 0 mmc 0
Signed-off-by: John Tobias <john.tobias.ph@gmail.com>
John Tobias [Thu, 7 Apr 2016 23:49:03 +0000 (16:49 -0700)]
USB: g_dnl: Change device class
The USB Mass Storage (ums) works in Windows, Linux and OS X (EL Capitan).
But, not in OS X (Yosemite). By applying the said patch, it extends
the ums support.
Signed-off-by: John Tobias <john.tobias.ph@gmail.com>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Test HW: Odroid XU3 (./test/py UMS + DFU tests)
Tested-by: John Tobias <john.tobias.ph@gmail.com>
Linux:
- Run ums to expose all my eMMC partition - shows all correctly
- Run ums to expose only 1 partition of my eMMC - show correctly
Windows:
- Run ums to expose all my eMMC partition - it detects but it prompts,
if I want to format it (due to a non windows partition)
- Run ums to expose only the FAT32 partition - it show the partition
correctly.
Stefan Roese [Mon, 18 Apr 2016 12:22:05 +0000 (14:22 +0200)]
arm: socfpga: socrates: Add eth0 alias to enable ethernet
This enables full ethernet usage, including U-Boot to write the board
specific MAC address (ethaddr) into the DT blob before passing it to
Linux.
Without this, the ethaddr is not detected in U-Boot at all, resulting
in this error upon bootup:
...
Model: EBV SOCrates
Net:
Error: ethernet@
ff702000 address not set.
No ethernet found.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Stefan Roese [Mon, 18 Apr 2016 12:22:04 +0000 (14:22 +0200)]
arm: socfpga: Fix typos in DT files (environmnet -> environment)
Fix a small typo in some of the SoCFPGA dts files that has spread via
copy-and-paste.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Tue, 5 Apr 2016 21:17:35 +0000 (23:17 +0200)]
ddr: altera: Repair DQ window centering code
The code uses a lot of signed numbers, which ended up in variables
of unsigned type, which resulted in all sorts of underflows. This
in turn caused incorrect calibration on certain boards. Moreover,
repair the readout of the DQ delay, which was being pulled from
wrong register.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Tue, 5 Apr 2016 09:18:38 +0000 (11:18 +0200)]
ddr: altera: Staticize global variables
Just staticize global variables in sequencer, since there is no
point in having these symbols available outside of the DDR code.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 19:21:05 +0000 (21:21 +0200)]
ddr: altera: Make DLEVEL behavior inclusive
Originally, the DLEVEL selects the debug level within the sequencer code,
but only displays the messages on that particular debug level. Tweak the
handling such that for particular debug level, debug messages on that
level and lower are displayed. This allows better regulation of debug
message verbosity.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 19:16:18 +0000 (21:16 +0200)]
ddr: altera: Zero DM IN delay in scc_mgr_zero_group()
This one last set of delay configuration registers was not properly
zeroed out originally, fix it and zero them out.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 17:10:12 +0000 (19:10 +0200)]
ddr: altera: Remove unnecessary ODT mode config
There is no point in resetting the ODT setting if the write test
failed, since the code will always retry the calibration and thus
reconfigure the ODT anyway OR the code will fail calibration and
halt.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 16:41:53 +0000 (18:41 +0200)]
ddr: altera: Remove unnecessary update of the SCC
Every invocation of the scc_mgr_set_dqs_en_delay_all_ranks() is
followed by SCC manager update. Moreover, only this function
triggers the SCC manager update internally. Thus, remove the
internal invocation to avoid triggering the update twice.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 15:52:21 +0000 (17:52 +0200)]
ddr: altera: Fix DRAM end value in protection rule
The hi address bitfield in the protection rule must be set to
the last address in the region which the rule represents. The
behavior is now in-line with code generated by Quartus 15.1 .
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 15:28:16 +0000 (17:28 +0200)]
ddr: altera: Fix scc_mgr_set() argument order
The code should be setting registers to zero, not one register to value.
Swap the order of arguments to correct the behavior. The behavior is now
in-line with code generated by Quartus 15.1 .
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Tue, 5 Apr 2016 21:41:56 +0000 (23:41 +0200)]
ddr: altera: Tweak DQS tracking enable handling
In the most unlikely case the DQS tracking was to be disabled,
make sure we do not errornously re-enable it. Note that DQS
tracking is enabled on all systems observed thus far.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Marek Vasut [Mon, 4 Apr 2016 14:07:11 +0000 (16:07 +0200)]
ddr: altera: Replace ad-hoc constant with macro
The bit 22 is in fact DQS tracking enable bit (dqstrken) and there
is a macro for this bit already, so use it.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Fabio Estevam [Mon, 18 Apr 2016 12:56:16 +0000 (09:56 -0300)]
Add initial support for Technexion's PICO-IMX6UL-EMMC board
Add support for Technexion's PICO-IMX6UL-EMMC board.
For information about this board, please visit:
http://www.technexion.com/products/pico/pico-som/pico-imx6-emmc
Signed-off-by: Richard Hu <richard.hu@technexion.com>
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Fabio Estevam [Mon, 18 Apr 2016 12:56:15 +0000 (09:56 -0300)]
MX6UL: Add definition for UART6 base address
Define the UART6_BASE_ADDR for MX6UL.
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Akshay Bhat [Tue, 12 Apr 2016 22:14:00 +0000 (18:14 -0400)]
board: ge: bx50v3: Fix to meet LVDS display power on timing
On a reset/reboot, the display power needs to be off for atleast 500ms
before turning it back on. So add a delay to the boot process to meet
the display timing requirement.
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Akshay Bhat [Tue, 12 Apr 2016 22:13:59 +0000 (18:13 -0400)]
board: ge: bx50v3: Use pwm for display backlight
Setup the LCD backlight brightness control pin to use PWM
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Akshay Bhat [Tue, 12 Apr 2016 22:13:58 +0000 (18:13 -0400)]
board: ge: bx50v3: Setup LDB_DI_CLK source
To generate accurate pixel clocks required by the displays we need to
set the ldb_di_clk source on bx50v3 to PLL3 and b850v3 to PLL5. Since
PLL5 is disabled on reset, we need to enable PLL5.
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Akshay Bhat [Tue, 12 Apr 2016 22:13:57 +0000 (18:13 -0400)]
board: ge: bx50v3: Split display setup function
B450v3/B650v3 uses single channel LVDS and does not support HDMI.
B850v3 uses dual channel LVDS and supports HDMI. Hence split the display
setup into two different functions.
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Akshay Bhat [Tue, 12 Apr 2016 22:13:56 +0000 (18:13 -0400)]
imx: mx6: Fix procedure to switch the parent of LDB_DI_CLK
Due to incorrect placement of the clock gate cell in the ldb_di[x]_clk tree,
the glitchy parent mux of ldb_di[x]_clk can cause a glitch to enter the
ldb_di_ipu_div divider. If the divider gets locked up, no ldb_di[x]_clk is
generated, and the LVDS display will hang when the ipu_di_clk is sourced from
ldb_di_clk.
To fix the problem, both the new and current parent of the ldb_di_clk should
be disabled before the switch. This patch ensures that correct steps are
followed when ldb_di_clk parent is switched in the beginning of boot.
This patch was ported from the 3.10.17 NXP kernel
http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/commit/?h=imx_3.10.17_1.0.1_ga&id=
eecbe9a52587cf9eec30132fb9b8a6761f3a1e6d
NXP errata number: ERR009219, EB821
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
Marek Vasut [Tue, 5 Apr 2016 08:54:06 +0000 (10:54 +0200)]
arm: mx5: Fix NAND image generation
The echo -ne "\xNN" does not work in certain bourne-compatible shells, like
dash. The recommended way of hex->char conversion is using printf(1), but
there is a pitfall here. The GNU printf does support "\xNN" format, but
according to the opengroup documentation, this is not part of POSIX. The
POSIX printf only defines "\NNN" where N is octal. Thus, for the sake of
compatibility, we use that.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
Justin Waters [Wed, 13 Apr 2016 21:03:18 +0000 (17:03 -0400)]
board: ge: bx50v3: Disable unused pins
Certain pins are not used on the i.MX6, and should have a neutral
pad configuration in order to reduce electrical interference on
the board. This commit defines these pins with a default value
rather than relying on the system defaults.
Signed-off-by: Justin Waters <justin.waters@timesys.com>
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Tom Rini [Tue, 19 Apr 2016 12:07:41 +0000 (08:07 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-mips
Purna Chandra Mandal [Mon, 18 Apr 2016 13:01:38 +0000 (18:31 +0530)]
MIPS: bootm: Add fixup of '/memory' node.
MIPS arch do not update 'reg' property of /memory node.
As a result Linux bootup will not work unless board.dts
file contains right /memory offset-size information or
board implements required memory fixup.
Fixing by renaming (unused) _arch_fixup_memory_node_ to
_arch_fixup_fdt_ in arch/mips/lib/bootm.c inline with ARM arch.
Signed-off-by: Purna Chandra Mandal <purna.mandal@microchip.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Robert P. J. Day [Sat, 16 Apr 2016 09:53:07 +0000 (05:53 -0400)]
Kconfig: Simple aesthetic/grammar fixes to top-level Kconfig
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Stephen Warren [Fri, 15 Apr 2016 20:40:37 +0000 (14:40 -0600)]
ARM: always perform per-CPU GIC init
The current code in ARMv8's lowlevel_init() skips the per-CPU GIC
initialization ifndef CONFIG_ARMV8_MULTIENTRY. However, the per-CPU init
should always occur; it's just the one-time init that should only happen
on the master.
Once this consideration is taken into account, the only difference between
the paths when CONFIG_ARMV8_MULTIENTRY is undefined/defined is the use of
branch_if_slave. Naively, any unified code would need to invoke this
conditionally upon ifdef CONFIG_ARMV8_MULTIENTRY. However, branch_if_slave
already checks CONFIG_ARMV8_MULTIENTRY and does nothing if it isn't
defined, so we don't even need that ifdef at the call site.
Reported-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Akshay Bhat [Thu, 14 Apr 2016 17:55:08 +0000 (13:55 -0400)]
bx50v3: Enable CONFIG_OF_LIBFDT in defconfig
As of commit
69e173eb57d1f4848f070c83456096ba5d2ba1b4, CONFIG_OF_LIBFDT
needs to be selected in defconfig instead of board specific header file.
Hence enable CONFIG_OF_LIBFDT in defconfig.
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Alexander Graf [Thu, 14 Apr 2016 14:07:54 +0000 (16:07 +0200)]
efi_loader: Fall back to fdtfile naming convention
When there is no $fdtfile variable set, we still have a good chance
that on 32bit arm the fdtfile really is just called $soc-$board.dtb.
Enable the exports for $soc and $board in our distr defaults and make
use of them in the efi boot script.
Reported-by: Andreas Faerber <afaerber@suse.de>
Reported-by: Stephen Warren <swarren@wwwdotorg.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Alexander Graf [Thu, 14 Apr 2016 14:07:53 +0000 (16:07 +0200)]
efi_loader: Pass fdt address directly to bootefi cmd
The bootefi cmd today fetches its device tree pointer from either the
location appointed by "fdt addr" with a fallback to the U-Boot control
fdt.
This integration is unusual for U-Boot and diverges from the way we
usually handle parameters to boot commands. So let's pass the fdt
directly into the bootefi command instead and move the control fdt
logic into the distro boot script.
Signed-off-by: Alexander Graf <agraf@suse.de>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Tom Rini [Mon, 18 Apr 2016 17:57:42 +0000 (13:57 -0400)]
sandbox: Set CONFIG_SYS_CPU
Give a valid value here as well to allow things which want
CONFIG_ENV_VARS_UBOOT_CONFIG to build
Signed-off-by: Tom Rini <trini@konsulko.com>
Robert P. J. Day [Thu, 14 Apr 2016 10:15:07 +0000 (06:15 -0400)]
board: Remove overlooked vestiges of "dave" board.
Apparently, all "dave"-related vendor content was removed in commit
5344cc1a82fcc2817d4671696b3939b0dfa4323e; remove remaining directory
board/dave/, which consists solely of board/dave/common/flash.c.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Tang Yuantian [Thu, 14 Apr 2016 08:21:00 +0000 (16:21 +0800)]
ahci: flush dcache before issuing command
Ensure data the following sata command used is flushed out of dcache
and written to physical memory or timeout error may happen.
Signed-off-by: Tang Yuantian <yuantian.tang@nxp.com>
Stephen Warren [Thu, 14 Apr 2016 04:29:52 +0000 (22:29 -0600)]
serial: bcm283x_mu: make pending values more explicit
dm_serial_ops.pending should return the number of characters, not just a
valid C Boolean integer value. The existing code does already does this,
but only as an accident since BCM283X_MU_LSR_RX_READY happens to be
BIT(0). Enhance the code to be more explicit about the values it returns.
Suggested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Robert P. J. Day [Wed, 13 Apr 2016 22:30:16 +0000 (18:30 -0400)]
doc: Updated README.ext4
Clean up the ext4 README file.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Robert P. J. Day [Wed, 13 Apr 2016 15:40:01 +0000 (11:40 -0400)]
Remove references to CONFIG_CMD_EXT3, no such command
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Andreas Färber [Wed, 13 Apr 2016 12:16:22 +0000 (14:16 +0200)]
dragonboard410c: Fix environment variables
Some variables for the distro boot commands were missing, using some
custom name instead. Rename them.
Cc: Mateusz Kulikowski <mateusz.kulikowski@gmail.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Andreas Färber [Wed, 13 Apr 2016 12:04:38 +0000 (14:04 +0200)]
efi_loader: Handle memory overflows
jetson-tk1 has 2 GB of RAM at 0x80000000, causing gd->ram_top to be zero.
Handle this by either avoiding ram_top or by using the same type as
ram_top to reverse the overflow effect.
Cc: Alexander Graf <agraf@suse.de>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Alexander Graf <agraf@suse.de>
Lokesh Vutla [Wed, 13 Apr 2016 04:27:04 +0000 (09:57 +0530)]
ARM: AM43xx: Fix BOOT_DEVICE_USB ID
commit
62c5674ea136a ("omap: SPL boot devices cleanup and completion")
cleans up the boot device ids for amx3xx soc. But mistakenly updates wrong
device IDs for AM43xx USB. Fixing the same here.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Lokesh Vutla [Wed, 13 Apr 2016 04:21:00 +0000 (09:51 +0530)]
configs: ks2: move CMD_NAND to defconfigs
NAND is not yet enabled on all Keystone2 platforms. So enabled
CMD_NAND in the respective defconfigs only if available.
Reported-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Lokesh Vutla [Wed, 13 Apr 2016 04:20:59 +0000 (09:50 +0530)]
memory: Move TI_AEMIF config to KCONFIG
Not all Keystone2 devices has AEMIF NAND controller. So adding Kconfig
entry for CONFIG_TI_AEMIF and enabling it in respective defconfigs on
platforms with AEMIF controller.
Reported-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Tom Rini [Tue, 12 Apr 2016 19:11:24 +0000 (15:11 -0400)]
drivers/power/pmic/pm8916.c: Make usid be uint32_t
If get_dev_addr fails it will return FDT_ADDR_T_NONE and:
>>> "priv->usid == 4294967295U" is always false regardless of the values of its operands. This occurs as the logical operand of if.
Cc: Mateusz Kulikowski <mateusz.kulikowski@gmail.com>
Reported-by: Coverity (CID: 143914)
Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Tue, 12 Apr 2016 19:11:23 +0000 (15:11 -0400)]
drivers/gpio/pm8916_gpio.c: Make pid be uint32_t
If get_dev_addr fails it will return FDT_ADDR_T_NONE and:
>>> "priv->pid == 4294967295U" is always false regardless of the values of its operands. This occurs as the logical operand of if.
Cc: Mateusz Kulikowski <mateusz.kulikowski@gmail.com>
Reported-by: Coverity (CID: 143913)
Signed-off-by: Tom Rini <trini@konsulko.com>
angelo@sysam.it [Mon, 11 Apr 2016 22:30:59 +0000 (00:30 +0200)]
m68k: fix broken buildman m68k
fix 19/48 broken board compilations, due to a now too smal 16-bit
relative jump
Signed-off-by: Angelo Dureghello <angelo@sysam.it>
Acked-by: Marek Vasut <marex@denx.de>
Acked-by: Heiko Schocher <hs@denx.de>
Eric Nelson [Mon, 11 Apr 2016 22:21:37 +0000 (15:21 -0700)]
gunzip.c: use block layer for writes
Call blk_dwrite to ensure that the block cache is notified
if enabled and remove build breakage when CONFIG_BLK is enabled.
Signed-off-by: Eric Nelson <eric@nelint.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Alexander Graf [Mon, 11 Apr 2016 21:51:02 +0000 (23:51 +0200)]
efi_loader: Expose ascending efi memory map
The EFI memory map does not need to be in a strict order, but 32bit
grub2 does expect it to be ascending. If it's not, it may try to
allocate memory inside the U-Boot data memory region.
We already sort the memory map in descending order, so let's just
reverse it when we pass it to a payload.
Signed-off-by: Alexander Graf <agraf@suse.de>
Tested-by: Andreas Färber <afaerber@suse.de>
Alexander Graf [Mon, 11 Apr 2016 21:51:01 +0000 (23:51 +0200)]
efi_loader: Put fdt into convenient location
The uEFI spec doesn't dictate where the device tree should live at, but
legacy 32bit ARM grub2 has some assumptions that it may stay at its place
when it's already loaded by the firmware.
So let's put it somewhere where Linux that comes after would happily find
it - around the recommended 128MB line.
Signed-off-by: Alexander Graf <agraf@suse.de>
Tested-by: Andreas Färber <afaerber@suse.de>
Alexander Graf [Mon, 11 Apr 2016 21:20:39 +0000 (23:20 +0200)]
efi_loader: Always flush in cache line size granularity
The cache line flush helpers only work properly when they get aligned
start and end addresses. Round our flush range to cache line size. It's
safe because we're guaranteed to flush within a single page which has the
same cache attributes.
Reported-by: Marek Vasut <marex@denx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Tested-by: Andreas Färber <afaerber@suse.de>
Mateusz Kulikowski [Mon, 11 Apr 2016 20:45:46 +0000 (22:45 +0200)]
dragonboard410c: Add CONFIG_SYS_CACHELINE_SIZE
Add missing define to board header file.
Signed-off-by: Mateusz Kulikowski <mateusz.kulikowski@gmail.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:55:26 +0000 (16:55 +0200)]
efi_loader: Use system fdt as fallback
When the user did not pass any device tree or the boot script
didn't find any, let's use the system device tree as last resort
to get something the payload (Linux) may understand.
This means that on systems that use the same device tree for U-Boot
and Linux we can just share it and there's no need to manually provide
a device tree in the target image.
While at it, also copy and pad the device tree by 64kb to give us
space for modifications.
Signed-off-by: Alexander Graf <agraf@suse.de>
Tested-by: Andreas Färber <afaerber@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:21 +0000 (16:16 +0200)]
distro: Enable iso partition code
Now that we can properly boot EFI payloads from iso el torito
images, let's enable support for isos by default in the distro
header.
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:20 +0000 (16:16 +0200)]
efi_loader: Increase path string to 32 characters
Whenever we want to tell our payload about a path, we limit ourselves
to a reasonable amount of characters. So far we only passed in device
names - exceeding 16 chars was unlikely there.
However by now we also pass real file path information, so let's increase
the limit to 32 characters. That way common paths like "boot/efi/bootaa64.efi"
fit just fine.
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:19 +0000 (16:16 +0200)]
efi_loader: Pass file path to payload
The payload gets information on where it got loaded from. This includes
the device as well as file path.
So far we've treated both as the same thing and always gave it the device
name. However, in some situations grub2 actually wants to find its loading
path to find its configuration file.
So let's split the two semantically separte bits into separate structs and
pass the loaded file name into our payload when we load it using "load".
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:18 +0000 (16:16 +0200)]
efi_loader: Add el torito support
When loading an el torito image, uEFI exposes said image as a raw
block device to the payload.
Let's do the same by creating new block devices with added offsets for
the respective el torito partitions.
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:17 +0000 (16:16 +0200)]
efi_loader: Split drive add into function
The snippet of code to add a drive to our drive list needs to
get called from 2 places in the future. Split it into a separate
function.
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:16 +0000 (16:16 +0200)]
iso: Allow 512 byte sector size
Real CD-ROMs are pretty obsolete these days. Usually people still keep
iso files around, but just put them on USB sticks or SD cards and expect
them to "just work".
To support this use case with El Torito images, add support for 512 byte
sector size to the iso parsing code.
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:15 +0000 (16:16 +0200)]
iso: Start with partition 1
The generic partition code treats partition 0 as "whole disk". So
we should start with partition 1 as the first partition in the iso
partition table.
Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 11 Apr 2016 14:16:14 +0000 (16:16 +0200)]
iso: Make little endian and 64bit safe
The iso partition table implementation has a few endian and 64bit
problems. Clean it up a bit to become endian and bitness safe.
Signed-off-by: Alexander Graf <agraf@suse.de>
Steve Kipisz [Fri, 8 Apr 2016 22:01:29 +0000 (17:01 -0500)]
board: ti: am57xx: Add support for am572x idk in SPL
The AM572x-IDK board (Industrial Dev Kit) is a board based on TI's AM5728x
SOC which has a dual core 1.5GHz A15 processor. This board is a development
platform for the Industrial market with:
- 2GB of DDR3L
- Dual 1Gbps Ethernet
- HDMI,
- PRU-ICSS
- uSD
- 16GB eMMC
- CAN
- RS-485
- PCIe
- USB3.0
- Video Input Port
- Industrial IO port and expansion connector
The link to the data sheet and TRM can be found here:
http://www.ti.com/product/AM5728
NOTE: DT support is still pending upstream kernel acceptance but we
should be able to get the base system support with this patch.
Signed-off-by: Schuyler Patton <spatton@ti.com>
Signed-off-by: Steve Kipisz <s-kipisz2@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Schuyler Patton [Fri, 8 Apr 2016 21:53:44 +0000 (16:53 -0500)]
board: ti: am57xx: Update EMIF SDRAM 1 and 3 Timings
Update EMIF data based on recommendations from the now standard TI
EMIF tool version 1.1.1 based on 256MBx16 DDR3L Kingston D2516EC4BXGGB
data sheet
Update T_RRD from 5 to 6 based on AM57xx TRM -
Minimum number of DDR cycles from activate to ativate for a different
bank, minus 1.
Update T_CKESR from 4 to 3 based on AM57xx TRM - Minimum number of DDR
clocks cycles for which SDRAM must remain in self refresh, minus 1.
Signed-off-by: Schuyler Patton <spatton@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Daniel Allred [Thu, 7 Apr 2016 02:59:34 +0000 (21:59 -0500)]
am57x: Move CONS_INDEX to Kconfig
- Move the CONS_INDEX selection out of CONFIG_SYS_EXTRA_OPTIONS and
into Kconfig proper.
- Edit the relevant am57x configs to remove the now unneeded
CONFIG_SYS_EXTRA_OPTIONS.
Signed-off-by: Daniel Allred <d-allred@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Michal Simek [Wed, 6 Apr 2016 18:28:04 +0000 (20:28 +0200)]
cmd: fdt: Use separate CMD_FDT Kconfig entry instead of OF_LIBFDT
Create CMD_FDT Kconfig entry to have an option to disable fdt command
which is not required for small configuration which requires libfdt
only.
Enable it by default for all targets which enables OF_LIBFDT.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
[trini: Fixup flea3/sandbox/id8313/siemens-am33xx/smartweb]
Signed-off-by: Tom Rini <trini@konsulko.com>
Michal Simek [Wed, 6 Apr 2016 18:28:03 +0000 (20:28 +0200)]
kconfig: Move CONFIG_OF_LIBFDT to Kconfig
This patch follows work done by:
"Move CONFIG_OF_LIBFDT to Kconfig"
(sha1:
69e173eb57d1f4848f070c83456096ba5d2ba1b4)
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
[trini: Add xpress* to the patch]
Signed-off-by: Tom Rini <trini@konsulko.com>
Steve Rae [Mon, 4 Apr 2016 19:59:45 +0000 (12:59 -0700)]
bcm281xx: save ENV to MMC
Enable saving ENV to MMC for the bcm281xx boards.
Signed-off-by: Steve Rae <srae@broadcom.com>
Steve Rae [Mon, 4 Apr 2016 19:59:44 +0000 (12:59 -0700)]
bcm281xx: enable CONFIG_OF_LIBFDT
set the Kconfig parameter: CONFIG_OF_LIBFDT
Signed-off-by: Steve Rae <srae@broadcom.com>
Mugunthan V N [Mon, 4 Apr 2016 11:58:05 +0000 (17:28 +0530)]
configs: dra74_evm: enable mmc driver model
enable mmc driver model for dra74 evm as omap_hsmmc
supports driver model
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Mugunthan V N [Mon, 4 Apr 2016 11:58:04 +0000 (17:28 +0530)]
configs: dra72_evm: enable mmc driver model
enable mmc driver model for dra72 evm as omap_hsmmc
supports driver model
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Mugunthan V N [Mon, 4 Apr 2016 11:58:03 +0000 (17:28 +0530)]
ARM: dts: dra7xx: am57xx: fix cd-gpios definition as per hardware design and dt binding docs
As per mmc device tree binding documentation card detect gpio has
to be active low signal. When a hardware is designed with active
high card detect, gpio polarity has to be changed with
cd-inverted dt property.
In DRA72x and AM57xx EVMs the card detect gpio is designed as
active low gpio. So correcting the dt card detect gpio
definition.
Also adding card-detect gpio for DRA74x EVM.
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Mugunthan V N [Mon, 4 Apr 2016 11:58:02 +0000 (17:28 +0530)]
ARM: dts: am43xx: fix cd-gpios definition as per hardware design and dt binding docs
As per mmc device tree binding documentation card detect gpio has
to be active low signal. When a hardware is designed with active
high card detect, gpio polarity has to be changed with
cd-inverted dt property.
In AM43xx the card detect gpio is designed as active low gpio.
So correcting the dt card detect gpio definition.
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Mugunthan V N [Mon, 4 Apr 2016 11:58:01 +0000 (17:28 +0530)]
drivers: mmc: omap_hsmmc: request cd and wp gpios when DM_MMC is defined
Add request gpio for CD and WP gpios, so that the gpio can be
used for the respective purposes.
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Mugunthan V N [Mon, 4 Apr 2016 09:52:49 +0000 (15:22 +0530)]
drivers: mmc: omap_hsmmc: Fix conversion of address to a pointer
omap_hsmmc driver directly typecasts fdt_addr_t to a pointer.
This is not strictly correct, as it gives a build warning when
fdt_addr_t is u64. So, use map_physmem for a proper typecasts.
This is inspired by commit
167efe01bc5a9 ("dm: ns16550: Use an address
instead of a pointer for the uart base")
drivers/mmc/omap_hsmmc.c: In function ‘omap_hsmmc_ofdata_to_platdata’:
drivers/mmc/omap_hsmmc.c:776:20: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
priv->base_addr = (struct hsmmc *)dev_get_addr(dev);
^
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Peng Fan [Sun, 3 Apr 2016 13:52:13 +0000 (21:52 +0800)]
common: env: support sata device
Introduce env support for sata device.
1. Implement write_env/read_env/env_relocate_spec/saveenv/sata_get_env_dev
2. If want to enable this feature, define CONFIG_ENV_IS_IN_SATA, and
define CONFIG_SYS_SATA_ENV_DEV or implement your own sata_get_ev_dev.
Signed-off-by: Peng Fan <van.freenix@gmail.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Heiko Schocher <hs@denx.de>
Cc: Stuart Longland <stuartl@vrt.com.au>
Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Cc: Tom Rini <trini@konsulko.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Stephen Warren [Sat, 16 Apr 2016 02:42:59 +0000 (20:42 -0600)]
ARM: rpi: fix 64-bit CONFIG_SYS_TEXT_BASE
The Pi firmware has changed the default "kernel" load address for 64-bit
mode. The authors have confirmed that this is a deliberate and long-term
change. Adapt U-Boot to the new value.
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Tom Rini [Tue, 12 Apr 2016 19:11:22 +0000 (15:11 -0400)]
drivers/tpm/tpm_tis_sandbox.c: Fix uninitialized variable use
In rollback_space_kernel we were not initializing the reserved fields
which should be for safety sake, and doing memset here means we don't
need to set the version field specifically either.
Reported-by: Coverity (CID: 143917)
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Stephen Warren [Wed, 6 Apr 2016 18:49:19 +0000 (12:49 -0600)]
fdt: implement dev_get_addr_name()
This function parses the reg property based on an index found in the
reg-names property. This is required for bindings that are written
using reg-names rather than hard-coding indices in reg.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tom Rini [Tue, 15 Mar 2016 17:20:23 +0000 (13:20 -0400)]
sandbox: Enable many more commands
- Set CONFIG_SYS_CACHELINE_SIZE to ARCH_DMA_MINALIGN as that should be
good enough.
- Make <asm/io.h> include <asm/types.h> like other arches do
- Enable many many more drivers in sandbox_defconfig so that we can get
more build-time testing on this platform.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tom Rini [Sun, 20 Mar 2016 14:10:28 +0000 (10:10 -0400)]
test/dm/core.c: Make pre-reloc test use pre-reloc struct
LLVM 3.5 noted:
test/dm/core.c:41:35: warning: unused variable 'test_pdata_pre_reloc' [-Wunused-const-variable]
static const struct dm_test_pdata test_pdata_pre_reloc = {
And the correct fix here is that the driver_info_pre_reloc test should
use the test_pdata_pre_reloc not test_pdata_manual variable
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Jagan Teki [Thu, 17 Mar 2016 06:53:19 +0000 (12:23 +0530)]
board: README.sandbox: Update dm test command
Update dm test command with pytest instead of ./test/dm/test-dm.sh
Cc: Tom Rini <trini@konsulko.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Jagan Teki <jteki@openedev.com>
Jagan Teki [Thu, 17 Mar 2016 06:53:18 +0000 (12:23 +0530)]
doc: driver-model: Update dm tests run using test.py
Since all the tests are implemented in pytest infrastructure,
So update the dm tests with the same instead of ./test/dm/test-dm.sh
Cc: Tom Rini <trini@konsulko.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Jagan Teki <jteki@openedev.com>
Stefan Roese [Wed, 16 Mar 2016 08:58:01 +0000 (09:58 +0100)]
dm: device.c: Minor coding-style fix
Fix multi-line comment indentation in device_bind()
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>