These functions tried to access two static tables before relocation
(board_early_init_f is executed before relocation). But these static
tables lie in the bss section which is not valid before relocation.
These accesses then overwrote some parts of u-boot binary before it was
relocated. For the kmnusa build, this results in a corrupted important
env variable (bootcmd) but it may be that some other parts of the u-boot
binary are corrupted.
This patch solves this problem by moving all the kw_gpio_* calls to
board_init, which should be early enough in the boot sequence. The only
calls that could not be moved is the one for the SOFT (bitbang) I2C, and
they have been replaced by a direct access to the GPIO dataout Control
register to set the two GPIOs as output.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
Add a function to read the dip_switch on kmcoge5un. If the
switch is set the actual_bank is set to 0 and this SW is
booted.
Signed-off-by: Thomas Herzmann <thomas.herzmann@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
The PCIe FPGAs now have to support 2 resets: one for the non traffic
affecting part (PCIe) and one for the traffic affecting part.
When the FPGA is not reconfigured, we only reset the PCIe part.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
In order to be able to perform board resets without interrupting the
traffic, the configuration of an already properly configured FPGA is
skipped.
This is because some PCIe FPGAs embed some other function that must
continue to work over reset.
It is then the responsibility of the application to trigger a
reconfiguration when needed. This is done by lowering the FPGA_INIT_B
pin for delaying the configuration to u-boot @ next reboot, and then
lower the FPGA_PROGRAM_B signal.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Some very similar #defines for reg addresses are used in a later patch
(managed_switch support for km_arm).
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
This adds a first support of the FPGA download for a PCIe FPGA based
on the BOCO2 CPLD.
This takes place in 3 steps, all done accessing the SPICTRL reg of the
BOCO2:
1) start the FPGA config with an access to the FPGA_PROG bit
2) later in the boot sequence, wait for the FPGA_DONE bit to toggle to 1
for the end of the FPGA configuration (with a timeout)
3) reset the FPGA
4) finally remove the access to its config EEPROM from the FPGA so that
the CPU can update the FPGA configuration when the kernel is running
The boards with a PCIe FPGA but without BOCO2 still are supported.
The config option name is CONFIG_KM_FPGA_CONFIG
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Gerlando Falauto <gerlando.falauto@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
The additional headerfile is unneeded here, we can use the generic
km_kirkwood.h instead. And we can use the better config option
KM_PIGGY4_88E6061 for the specific features for boards with this
design in km_arm.c.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Gerlando Falauto <gerlando.falauto@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
Use the generic header km_kirkwood.h and get rid of the
board specific header.
changes for v2: rebased because of changes in other patches
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Gerlando Falauto <gerlando.falauto@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
For u-boot this board is similar to mgcoge3un. But some differences
are present. We have a different SDRAM on it and therefore a new
SDRAM config file. Additionaly this board has a direct MAC/MAC
connection from the kirkwood to a marvell simple switch without a
phy inbetween, this needs a new configuration for the mvgbe driver.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Gerlando Falauto <gerlando.falauto@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
This board is similar to portl2, but it has the u-boot environment
in a SPI NOR flash and not in an i2c eeprom like portl2 have.
Some other details:
- IVM EEPROM is at adress: pca9547:70:9
- PCI is enabled
- PIGGY4 is connected via MV88E6352 simple switch. There is no phy
between the simple switch and the kirkwood.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Gerlando Falauto <gerlando.falauto@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
We overwrite these weak functions from the kirkwood spi code to
use our own method to be able to switch between the SPI NOR and
the NAND flash. This is needed e.g. to update the u-boot. The former
command do_spi_toggle can therefore be removed. And the usage of
this command is removed from the u-boot update command in the
u-boot environment.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Prafulla Wadaskar <prafulla@marvell.com>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
With the new second save argument introduced by the previous patch, all
the calls to the function had to be fixed.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
The bootcounter (stored in the RAM) is not enough protected with the 4 Bytes
BOOTCOUNT_MAGIC against bit errors due to short power loss or holding a system
in RESET. It has been seen, that the bootcounter value has been changed due to
a bit flip on a system holding in RESET, but the BOOTCOUNT_MAGIC was still valid.
A bit pattern with 4000 bytes (after BOOTCOUNT_MAGIC) has been implemented,
which should be enough to detect a bit error.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
The Xilinx FPGA must be reconfigured each time the unit
reboots. The FPGA is connected to the GPIO pin 39 from kirkwood.
This patch triggers this pin for km_kirkwood_pci targets.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
We got dropped from Linux mach-types.h because of a lack of mainline
support and this is needed since the last Linux/u-boot mach-types
synchro.
This patch also defines CONFIG_MACH_TYPE for all keymile boards, as
this is a mandatory CONFIG for ARM boards now. The initialization
of gd->bd->bi_arch_number is removed form km_arm.c, our board file.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Prafulla Wadaskar <prafulla@marvell.com>
The mgcoge3un waits to be released from mgcoge3ne at startup.
This patch enhances this feature with the possibility to interrupt
this wait if a key is pressed.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
For files like the drivers/serial/serial.c, it must include the
platform file, as the CONFIG_SYS_NS16550_COM1 must reference to the
definition in the platform definition files.
Include the platform definition file in the config file, so that it
would decouple the dependence for the driver files.
Signed-off-by: Lei Wen <leiwen@marvell.com>
This reverts commit a2da616311.
THis was applied by accident - a more recent version of this change
was already present, see commit
9400f8f 2011-10-05 22:03:11 +0200 km_arm: enable POST for these boards
Signed-off-by: Wolfgang Denk <wd@denx.de>
The current km_arm boards have a Power-On test jumper. When this
jumper is set, this triggers some Power-On tests on the board.
This patch enables the support of this jumper for starting the
memory_regions test when the jumper is set.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Prafulla Wadaskar <prafulla@marvell.com>
The top level Makefile does not do any recursion into subdirs when
cleaning, so these clean/distclean targets in random arch/board dirs
never get used. Punt them all.
MAKEALL didn't report any errors related to this that I could see.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The current km_arm boards have a Power-On test jumper. When this
jumper is set, this triggers some Power-On tests on the board.
This patch enables the support of this jumper for starting the
memory_regions test when the jumper is set.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Prafulla Wadaskar <prafulla@marvell.com>
The file common.c grows in the past. So move the IVM specific
code into an ivm.c file.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Heiko Schocher <hs@denx.de>
cc: Wolfgang Denk <wd@denx.de>
Checkpatch.pl complains about the volatile qualifier in calls to
get_ram_size(). Remove this qualifier in the prototype and in the
calls where it is useless, and leave it only in the function body
where it is needed.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
This adds support for the keymile Kirkwood BEC portl2 board. This board
relies on the km_arm (km_kirkwood) BEC.
The egiga driver is configured for a 100M full-duplex, A/N off connnection
to the backplane. This board has always ethernet present, because it is
connected to the marvell switch similar to mgcoge3un. The reset_phy
functionality is also the same to mgcoge3un.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
cc: Heiko Schocher <hs@denx.de>
The phy is also configured with "RGMII clock transitions when data
stable" and "Class A driver for the direct backplane connection".
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
cc: Heiko Schocher <hs@denx.de>
We change default settings for egiga on mgcoge3un.
The reason we need this is that we have the gig port on mgcoge3un
connected using a back-to-back pair of PHYs. There are no magnetics and
because of that the port has to be run with a fixd configuration and
auto-negotiation must be disabled. In the default mode the egiga driver
uses autoneg to determine port speed - which defaults to 1G (we need
100M full duplex).
Add wait for the GPIO line connected to mgcoge3ne before
starting mgcoge3un. A board specific ethernet present function
was added, because on this board ethernet is always present.
The BOCO FPGA access was enhanced and changed to use register
definitions.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Acked-by: Heiko Schocher <hs@denx.de>
cc: Wolfgang Denk <wd@denx.de>
cc: Detlev Zundel <dzu@denx.de>
cc: Prafulla Wadaskar <prafulla@marvell.com>
The mgcoge2un target was only an intermediate step to mgcoge3un.
For this reason the mgcoge2un support was moved to mgcoge3un,
because it isn't needed to support both targets.
We add the BootROM init file for the mgcoge3un memphis RAM.
We also move the suen3 and suen8 boards into the correct category
in the MAINTAINERS file.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Acked-by: Heiko Schocher <hs@denx.de>
cc: Wolfgang Denk <wd@denx.de>
cc: Detlev Zundel <dzu@denx.de>
cc: Prafulla Wadaskar <prafulla@marvell.com>
This environment variable is used to set the bootcount address
for the kernel.
last_stage_init is not available for arm platforms. So the
calls to set_km_var and set_bootcount_addr are done in
misc_init_r.
Additionally some unneeded printouts were removed.
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Acked-by: Heiko Schocher <hs@denx.de>
cc: Wolfgang Denk <wd@denx.de>
cc: Detlev Zundel <dzu@denx.de>
cc: Prafulla Wadaskar <prafulla@marvell.com>
This patch adds last_stage_init to all keymile boards. And
in the last stage init some environment variables for u-boot
were set. Currently these are pnvramaddr, pram and var address.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Heiko Schocher <hs@denx.de>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
cc: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Wolfgang Denk <wd@denx.de>
cc: Kim Phillips <kim.phillips@freescale.com>
define KM_IVM_BUS and KM_ENV_BUS macros
KM_IVM_BUS is used to define the EEprom_ivm environment variable.
These macros allow the reuse of these I2C addresses in other code
locations.
remove unneeded code
On first HW versions the BOCO FPGA was behind a MUX device. These
HW versions are not supported anymore. And therefore this code can
be removed.
added LED initialization for SUEN3
The bootstat LED required to be initialized so to have a green
colour after start-up.
define CONFIG_SYS_TEXT_BASE
This is needed by the relocation code and is not the same for
our ARM BEC and thus needs to be defined.
remove memsize variable
An environment variable for memsize is not needed.
this can be get via the board info struct.
remove unneeded double access to bi_dram[i].size field
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Luca Haab <luca.haab@keymile.com>
Signed-off-by: Heiko Schocher <hs@denx.de>
cc: Prafulla Wadaskar <prafulla@marvell.com>
These new values are:
- enables UART0 and UART1 pins in MPP
- define some L2 cache settings
- changes a SDRAM timing to better fit the hardware
- removed three writes that were the same as the reset values
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Heiko Schocher <hs@denx.de>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
This patch renames the suen3 defines and functions to KM_KIRKWOOD
which is more generic and more precise, because these values
and functions where used by all suenX boards and not only suen3.
Signed-off-by: Lukas Roggli <lukas.roggli@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Heiko Schocher <hs@denx.de>
cc: Prafulla Wadaskar <prafulla@marvell.com>
- use I/O accessors
-> For accessing the FPGA therefore a struct km_bec_fpga
is introduced.
- no longer externs needed
- to defines, that only select functions, don;t assign a
numeric value
- Codingstyle changes to prevent checkpatch errors/warnings
Signed-off-by: Heiko Schocher <hs@denx.de>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
cc: Kim Phillips <kim.phillips@freescale.com>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Wolfgang Denk <wd@denx.de>
cc: Detlev Zundel <dzu@denx.de>
cc: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
The MACH_TYPE SUEN3 is now to specific for keymile boards, because
other boards similar to suen3 will follow. So the MACH_SUEN3 was renamed
to MACH_KM_KIRKWOOD.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Heiko Schocher <hs@denx.de>
cc: Wolfgang Denk <wd@denx.de>
cc: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
On first HW versions the BOCO FPGA was behind a MUX device. These
HW versions are not supported anymore. And therefore this code can
be removed, it is already unused.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Before this commit, weak symbols were not overridden by non-weak symbols
found in archive libraries when linking with recent versions of
binutils. As stated in the System V ABI, "the link editor does not
extract archive members to resolve undefined weak symbols".
This commit changes all Makefiles to use partial linking (ld -r) instead
of creating library archives, which forces all symbols to participate in
linking, allowing non-weak symbols to override weak symbols as intended.
This approach is also used by Linux, from which the gmake function
cmd_link_o_target (defined in config.mk and used in all Makefiles) is
inspired.
The name of each former library archive is preserved except for
extensions which change from ".a" to ".o". This commit updates
references accordingly where needed, in particular in some linker
scripts.
This commit reveals board configurations that exclude some features but
include source files that depend these disabled features in the build,
resulting in undefined symbols. Known such cases include:
- disabling CMD_NET but not CMD_NFS;
- enabling CONFIG_OF_LIBFDT but not CONFIG_QE.
Signed-off-by: Sebastien Carlier <sebastien.carlier@gmail.com>
When this define was introduced, the idea was to provide a soft
migration path for ARM boards to get adapted to the new relocation
support. However, other recent changes led to a different
implementation (ELF relocation), where this no longer works. By now
CONFIG_SYS_ARM_WITHOUT_RELOC does not only not help any more, but it
actually hurts because it obfuscates the actual code by sprinkling it
with lots of dead and non-working debris.
So let's make a clean cut and drop CONFIG_SYS_ARM_WITHOUT_RELOC.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Tested-by: Heiko Schocher <hs@denx.de>
Tested-by: Reinhard Meyer <u-boot@emk-elektronik.de>
After moving the definition of CONFIG_SYS_TEXT_BASE to the respective
board config files, all Marvell kirkwood board have just a single and
common entry in their config.mk files:
KWD_CONFIG = $(SRCTREE)/board/$(BOARDDIR)/kwbimage.cfg
Replace the only reference to KWD_CONFIG in the top level Makefile by
an equivalent setting, and remove all kirkwood config.mk files.
Signed-off-by: Wolfgang Denk <wd at denx.de>
Cc: Prafulla Wadaskar <prafulla at marvell.com>
Cc: Siddarth Gore <gores at marvell.com>
Cc: Simon Kagstrom <simon.kagstrom at netinsight.net>
Cc: Heiko Schocher <hs at denx.de>
Cc: Eric Cooper <ecc at cmu.edu>
Acked-by: Wolfgang Denk <wd at denx.de>
Signed-off-by: Prafulla Wadaskar <prafulla@marvell.com>
- use internal 2k security SRAM as RAM for early stack.
- do early inits in board_init_f()
Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Prafulla Wadaskar <prafulla@marvell.com>
The change is currently needed to be able to remove the board
configuration scripting from the top level Makefile and replace it by
a simple, table driven script.
Moving this configuration setting into the "CONFIG_*" name space is
also desirable because it is needed if we ever should move forward to
a Kconfig driven configuration system.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Change the implementation for arm926 to relocate the code to
an arbitrary address in RAM.
Adapt the TX25 (i.MX25), magnesium board to test the changes.
On the tx25 board TEXT_BASE is set to the final relocation
address to prevent one more copying of u-boot code
when relocating. More info see:
doc/README.arm-relocation
da850 board:
Tested-by: Ben Gardiner <bengardiner@nanometrics.ca>
Portions of this work were supported by funding from
the CE Linux Forum.
Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Ben Gardiner <bengardiner@nanometrics.ca>
Lots of code use this construct:
cmd_usage(cmdtp);
return 1;
Change cmd_usage() let it return 1 - then we can replace all these
ocurrances by
return cmd_usage(cmdtp);
This fixes a few places with incorrect return code handling, too.
Signed-off-by: Wolfgang Denk <wd@denx.de>
The hush shell dynamically allocates (and re-allocates) memory for the
argument strings in the "char *argv[]" argument vector passed to
commands. Any code that modifies these pointers will cause serious
corruption of the malloc data structures and crash U-Boot, so make
sure the compiler can check that no such modifications are being done
by changing the code into "char * const argv[]".
This modification is the result of debugging a strange crash caused
after adding a new command, which used the following argument
processing code which has been working perfectly fine in all Unix
systems since version 6 - but not so in U-Boot:
int main (int argc, char **argv)
{
while (--argc > 0 && **++argv == '-') {
/* ====> */ while (*++*argv) {
switch (**argv) {
case 'd':
debug++;
break;
...
default:
usage ();
}
}
}
...
}
The line marked "====>" will corrupt the malloc data structures and
usually cause U-Boot to crash when the next command gets executed by
the shell. With the modification, the compiler will prevent this with
an
error: increment of read-only location '*argv'
N.B.: The code above can be trivially rewritten like this:
while (--argc > 0 && **++argv == '-') {
char *arg = *argv;
while (*++arg) {
switch (*arg) {
...
Signed-off-by: Wolfgang Denk <wd@denx.de>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Add support for the ARM part of the mgcoge2, named suen3.
This board is based on the Marvell Kirkwood (88F6281) SoC.
As there come more board variants, common config options
are collected in include/configs/km_arm.h. Also, this board
use common code for all keymile boards, which is stored in
board/keymile/common/common.c
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Heiko Schocher <hs@denx.de>
