Lukasz Majewski [Fri, 1 Mar 2013 14:30:18 +0000 (15:30 +0100)]
usb:composite:fix Provide function data when addressing device with only one interface
This commit fixes problems with some non-standard requests send with
device address instead of interface address (bmRequestType.Receipent field).
This happens with dfu-util (debian version: 0.5), which address non standard
requests (like w_value=0x21 and bRequest=GET_DESCRIPTOR) to device.
Without this fix, the above request is STALLED, and hence causes dfu-util
to assume some standard configuration (packet size = 1024B instead of 4096B)
In turn it displays following errors:
Error obtaining DFU functional descriptor
Warning: Assuming DFU version 1.0
Warning: Transfer size can not be detected
...
Warning: Trying default transfer size 1024
This fix allows passing non-standard request to function setup code, where
it shall be handled.
Tested at: Trats (exynos4210)
Tested with:DFU and UMS gadgets
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Simon Glass [Tue, 5 Mar 2013 14:40:04 +0000 (14:40 +0000)]
sparc: Fix build warnings in serial.c
These macros are already defined in io.h so should not be declared in
serial.c.
serial.c:38:0: warning: "READ_BYTE" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:36:0: note: this is the location of the previous definition
serial.c:39:0: warning: "READ_HWORD" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:37:0: note: this is the location of the previous definition
serial.c:40:0: warning: "READ_WORD" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:38:0: note: this is the location of the previous definition
serial.c:41:0: warning: "READ_DWORD" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:39:0: note: this is the location of the previous definition
Signed-off-by: Simon Glass <sjg@chromium.org> Signed-off-by: Simon Glass <sjg@chromium.org>
Simon Glass [Tue, 5 Mar 2013 14:40:03 +0000 (14:40 +0000)]
sparc: Fix out-of-tree building
It isn't clear why the sparc cpu Makefile has its own compile line, but
it does not work correctly with an out-of-tree build. Removing it fixes
this problem. Perhaps it does not introduce others.
Simon Glass [Tue, 5 Mar 2013 14:39:52 +0000 (14:39 +0000)]
x86: Adjust board_f.c for x86
For x86, things have adjusted somewhat since this series was originally
written. It has its own way of running through initcalls which is actually
nicer than others archs.
Unfortunately this does introduce exceptions. We will soon require use of
generic board on x86, but until then we need to fit in with what is there,
and treat x86 as a special case.
Simon Glass [Mon, 11 Mar 2013 07:06:48 +0000 (07:06 +0000)]
Add CONFIG_SYS_SYM_OFFSETS to support offset symbols
Link symbols as created by the link script can either be absolute or
relative to the text start. This option switches between the two options
so that we can support both.
As we convert architectures over to generic board, we can see if this
option is actually needed, or whether it is possible to unify this feature
also.
Simon Glass [Tue, 5 Mar 2013 14:39:46 +0000 (14:39 +0000)]
arm: Remove use of board_early_init_r/last_stage_init()
These boards define CONFIG_LAST_STAGE_INIT and CONFIG_BOARD_EARLY_INIT_R
but these options are not available on ARM. Move them into the powerpc
common file instead.
This change affects: km_kirkwood_pci, mgcoge3un, kmnusa, kmcoge5un,
km_kirkwood and portl2.
Simon Glass [Tue, 5 Mar 2013 14:39:42 +0000 (14:39 +0000)]
Declare watchdog functions in watchdog.h
These functions are only available for powerpc and are not declared in a
header file. We want to use the rest function in two places (board_f and
board_r), so declare the functions in watchdog.h.
Simon Glass [Fri, 8 Mar 2013 13:45:27 +0000 (13:45 +0000)]
__HAVE_ARCH_GENERIC_BOARD controls availabilty of generic board
We are introducing a new unified board setup. Add a check to make sure that
board config files do not define CONFIG_SYS_GENERIC_BOARD unless their
architecture defines __HAVE_ARCH_GENERIC_BOARD
__HAVE_ARCH_GENERIC_BOARD will currently not be the default setting, but
we can switch this later when most architecture support generic board.
Simon Glass [Tue, 5 Mar 2013 14:39:40 +0000 (14:39 +0000)]
Introduce a basic initcall implementation
This library supports calling a list of functions one after the
other.
It is intended that we move to a more powerful initcall implementation
as proposed by Graeme Russ <graeme.russ@gmail.com>. For now, this allows
us to do the basics.
Simon Glass [Tue, 5 Mar 2013 14:39:37 +0000 (14:39 +0000)]
Introduce generic link section.h symbol files
We create a separate header file for link symbols defined by the link
scripts. It is helpful to have these all in one place and try to
make them common across architectures. Since Linux already has a similar
file, we bring this in even though many of the symbols there are not
relevant to us.
Each architecture has its own asm/sections.h where symbols specifc to
that architecture can be added. For now everything except AVR32 just
includes the generic header.
One change is needed in arch/avr32/lib/board.c to make this conversion
work.
Reviewed-by: Tom Rini <trini@ti.com> (version 5) Signed-off-by: Simon Glass <sjg@chromium.org>
Simon Glass [Tue, 5 Mar 2013 14:39:35 +0000 (14:39 +0000)]
Introduce generic u-boot.h file
This file holds the board info structure. We need this to be generic
for the unified board series, so create a structure which contains
the basic fields required by the main architectures.
1) Write the correct M/T Stop value to I2CSTAT after i2c write.
According to the spec, after finish the data transmission, we should
write a M/T Stop (I2C_MODE_MT | I2C_TXRX_ENA) to I2CSTAT instead of
a M/R Stop (I2C_MODE_MR | I2C_TXRX_ENA).
2) Not split the write to I2CSTAT into 2 steps in i2c read.
According to the spec, we should write the combined M/R Start value to
I2CSTAT after setting the slave address to I2CDS
3) Fix the mistake of making an equality check to an assignment.
In the case of I2C write with the zero-length address, while tranfering the
data, it should be an equality check (==) instead of an assignment (=).
Signed-off-by: Tom Wai-Hong Tam <waihong@chromium.org> Signed-off-by: Rajeshwari Shinde <rajeshwari.s@samsung.com>
Joe Hershberger [Fri, 8 Feb 2013 10:28:00 +0000 (10:28 +0000)]
CONFIG_BOOTDELAY default should not affect runtime
Because the code that handles bootdelay is compiled in conditionally
based on the default value, you are restricted in the default,
regardless of what you want the runtime options to be.
Change the source to always check if any default is given so that other
values can be selected and used at runtime.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Wolfgang Denk [Wed, 20 Feb 2013 04:53:16 +0000 (04:53 +0000)]
env: fix "env ask" command
The "env ask" traditionally uses a somewhat awkward syntax:
env ask name [message ...] [size]
So far, when a mesage was given, you always also had to enter a size.
If you forgot to do that, the command would terminate without any
indication of the problem.
To avoid incompatible changes of the interface, we now check the last
argument if it can be converted into a decimal number. If this is the
case, we assume it is a size; otherwise we treat it as part of the
message.
Also, add a space after the message fore easier reading,
and clean up help mesage.
Tom Rini [Thu, 21 Feb 2013 06:55:40 +0000 (06:55 +0000)]
cmd_fat.c: Note in fatread help about alignment requirements
When using the partial read feature of fatwrite the buffer we read into
can become unaligned not just due to initial location but the size of
our partial reads as well. Make this clear in the help text.
Stephen Warren [Thu, 28 Feb 2013 15:05:34 +0000 (15:05 +0000)]
cmd_part: don't print cmd name twice in help
The core implementation of "help" already prints the command name before
the help text of a specific command. Remove it from part's own help text
to avoid it being printed twice:
Tegra114 (Dalmore) # help part
part - disk partition related commands
Usage:
part part uuid <interface> <dev>:<part>
- print partition UUID
...
Signed-off-by: Stephen Warren <swarren@nvidia.com>
ps on BSD hosts (like OS X) do not provide the --no-headers switch nor
understand the AIX format descriptions. Unfortunately there seems no solution to
get the PIDs of children in a platfrom independent manner.
Therefore detect the OS and decide upon that which way to go.
This patch makes the MAKEALL script cleanly stoppable on bare OS X when using
the parallel builds of targets.
Additionally this patch removes double call to grep by a single call to sed for
GNU style child PID detection.
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com> Cc: Joe Hershberger <joe.hershberger@ni.com> Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Wolfgang Denk [Fri, 8 Mar 2013 10:51:32 +0000 (10:51 +0000)]
Feature Removal: disable "mtest" command by default
The "mtest" command is of little practical use (if any), and
experience has shown that a large number of board configurations
define useless or even dangerous start and end addresses. If not even
the board maintainers are able to figure out which memory range can be
reliably tested, how can we expect such from the end users? As this
problem comes up repeatedly, we rather do not enable this command by
default, so only people who know what they are doing will be
confronted with it.
As this changes the user interface, we allow for a grace period
before this change takes effect. For now, we make "mtest"
configurable through the CONFIG_CMD_MEMTEST variable, which is defined
in include/config_cmd_default.h; we also add an entry to
doc/feature-removal-schedule.txt which announces the removal of this
default setting in two releases from now, i. e. with v2013.07.
Signed-off-by: Wolfgang Denk <wd@denx.de> Cc: Tom Rini <trini@ti.com>
mpc512x: Adjust the DRAM init sequence to the datasheet spec
Do maintain a 200 usecs period of stable power and clock before
asserting the CKE signal and sending commands, have at least 200
DRAM clock cycles pass after initialization before data access.
Add new mpc5121e based ac14xx board and a new pinmux config
function for setting individual pinmux bit groups. This
function is used in ac14xx board code.
Remove CSx configurations from board code and only define
required CSx macros in the board config file to configure
chip select windows and parameters.
Signed-off-by: Anatolij Gustschin <agust@denx.de> Cc: Reinhard Arlt <reinhard.arlt@esd-electronics.com> Cc: Wolfgang Denk <wd@denx.de>
mpc512x: add common LAW and Chip Select configuration
Currently each mpc512x board has its own LAW and CS config code,
we should avoid this code duplication. Allow all boards to use
common code by only defining LAW and CS config macros like
CONFIG_SYS_CSx_START, CONFIG_SYS_CSx_SIZE and CONFIG_SYS_CSx_CFG.
Also allow common configuration of additional CS parameters by
CONFIG_SYS_CS_ALETIMING, CONFIG_SYS_CS_BURST, CONFIG_SYS_CS_DEADCYCLE
and CONFIG_SYS_CS_HOLDCYCLE options.
Stefan Roese [Thu, 7 Feb 2013 02:10:28 +0000 (02:10 +0000)]
mpc5200: a4m2k: Implement custom "dynamic" watchdog support
This patch adds a custom U-Boot command "wdogtoggle" which enables the
external hardware watchdog toggling via an GPIO pin on the a4m2k
board. After issuing this commands, the watchdog will be serviced
by U-Boot so that the user can use all U-Boot commands from the
prompt.
Stefan Roese [Thu, 7 Feb 2013 02:10:11 +0000 (02:10 +0000)]
mpc5200: Add a4m2k board port
This patch adds the a4m2k MPC5200B board port. Its a derivate of
the a3m071 board with only minor changes.
Additionally this patch includes some clean-up changes:
- Remove I2C support from a3m071 as its unused
- Fix/enhance default env variables
- Fix some comments
- Add newly introduced CONFIG_SPL_TARGET to automatically build
"u-boot-img.bin"
- Fix dtb patching in READ desciption for SPL Linux booting:
"fdt chosen" needs to get called to patch/create the chosen node.
- Add missing call to spl_board_init():
Define CONFIG_SPL_BOARD_INIT so that spl_board_init() will get
called in the SPL version.
Stefan Roese [Thu, 6 Dec 2012 03:41:57 +0000 (03:41 +0000)]
mpc5200: spl_boot.c: Change init oder to first enable printf
On MPC5200, the initial RAM (and gd) is located in the internal
SRAM. So we can actually call the preloader console init code
before calling initdram(). This makes serial output (printf)
available very early, even before SDRAM init, which has been
an U-Boot priciple from day 1.
Simon Glass [Thu, 28 Feb 2013 19:26:17 +0000 (19:26 +0000)]
x86: Adjust link device tree include file
This is currently set to coreboot.dtsi, but we cannot support this on
old device tree compilers (dtc <= 1.3), so adjust to use ARCH_CPU_DTS
to let the Makefile preprocessor sort this out.
Simon Glass [Thu, 28 Feb 2013 19:26:16 +0000 (19:26 +0000)]
x86: Add error checking to x86 relocation code
This does not actually change normal behaviour, but adds a check that
should detect corruption of relocation data (e.g. by using BSS data
prior to relocation).
Also add additional debugging output when enabled.
During this investigation, two situations have been seen:
1. calculate_relocation_address():
uintptr_t size = (uintptr_t)&__bss_end - (uintptr_t)&__text_start;
turns into 111166f: b8 83 c4 17 01 mov $0x117c483,%eax
Simon Glass [Thu, 28 Feb 2013 19:26:10 +0000 (19:26 +0000)]
x86: Add function to get top of usable ram
The memory layout calculations are done in calculate_relocation_address(),
and coreboot has its own version of this function. But in fact all we
really need is to set the top of usable RAM, and then the base version
will work as is.
So instead of allowing the whole calculate_relocation_address() function
to be replaced, create board_get_usable_ram_top() which can be used by
a board to specify the top of the area where U-Boot relocations to.
bootm: fix conditional compilation for bootm ramdisk subcommand
All code related to the bootm ramdisk subcommand is conditionally
enabled by CONFIG_SYS_BOOT_RAMDISK_HIGH except for the help message.
Replace the CONFIG_ARCH defines by CONFIG_SYS_BOOT_RAMDISK_HIGH
to fix this.
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Joe Hershberger [Thu, 1 Nov 2012 16:54:18 +0000 (16:54 +0000)]
ubifs: Allow ubifsmount volume reference by number
UBI can mount volumes by name or number The current code forces you
to name the volume by prepending every name with "ubi:".
>From fs/ubifs/super.c
* There are several ways to specify UBI volumes when mounting UBIFS:
* o ubiX_Y - UBI device number X, volume Y;
* o ubiY - UBI device number 0, volume Y;
* o ubiX:NAME - mount UBI device X, volume with name NAME;
* o ubi:NAME - mount UBI device 0, volume with name NAME.
Now any name passed in any of the above forms are allowed.
Also update the configs that referenced ubifsmount.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Simon Glass [Wed, 26 Dec 2012 09:53:36 +0000 (09:53 +0000)]
sandbox: Add 'sb' command to access filesystem features
The new 'sb' command is intended to deal with sandbox-specific features
that have no parallel in other archs. This commit adds two sub-commands
to list a directory and read a file from the host filesystem.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@ti.com>
Simon Glass [Wed, 26 Dec 2012 09:53:33 +0000 (09:53 +0000)]
fs: Move ls and read methods into ext4, fat
It doesn't make a lot of sense to have these methods in fs.c. They are
filesystem-specific, not generic code. Add each to the relevant
filesystem and remove the associated #ifdefs in fs.c.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@ti.com>
Simon Glass [Wed, 26 Dec 2012 09:53:31 +0000 (09:53 +0000)]
fs: Tell probe functions where to put their results
Rather than rely on global variables for the probe functions, pass in
the information that we need filled in. This allows us to potentially
keep the variables private to fs.c in the future, and the meaning of
the probe function is clearer.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@ti.com>
Simon Glass [Wed, 26 Dec 2012 09:53:30 +0000 (09:53 +0000)]
fs: Use filesystem methods instead of switch()
We can use the available methods and avoid using switch(). When the
filesystem is not supported, we fall through to the 'unsupported'
methods: fs_probe_unsupported() prints an error, so the others do
not need to.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@ti.com>
Simon Glass [Wed, 26 Dec 2012 09:53:29 +0000 (09:53 +0000)]
fs: Fully populate the filesystem method struct
There is a structure in fs.c with just a probe method. By adding methods
for other operations, we can avoid lots of #ifdefs and switch()s. As a
first step, create the structure ready for use.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@ti.com>