Lothar Waßmann [Mon, 24 Jun 2013 09:44:57 +0000 (11:44 +0200)]
am33xx: remove global variable gpmc_cfg
SPL and regular code use different addresses for global variables.
Thus the initial value from SPL code is not available in regular code.
Make the variable local to each file where it is used.
Lothar Waßmann [Mon, 24 Jun 2013 09:30:28 +0000 (11:30 +0200)]
am33xx: tx48: config cleanup
- whitespace cleanups
- remove CONFIG_SYS_SDRAM_CLK (use CONFIG_SYS_DDR_CLK from boards.cfg instead)
- remove dead code
- set cpu_clk ENV setting to CONFIG_SYS_MPU_CLK
- explicitly define the BBT area as partition and give userfs a fixed size
- add CONFIG_SPL_NAND_DRIVERS, remove CONFIG_SPL_NAND_SIMPLE
- fix whitespace error in TX48_MTDPARTS_CMD
Lothar Waßmann [Mon, 24 Jun 2013 09:13:29 +0000 (11:13 +0200)]
am33xx: tx48: general cleanup
- add timeout to polling loop for UART_RESETDONE
- setup gd for SPL
- stub out tx48_wdog_disable() if CONFIG_HW_WATCHDOG is not set
- make enable_*_pin_mux() static
Lothar Waßmann [Mon, 24 Jun 2013 09:07:51 +0000 (11:07 +0200)]
am33xx: general clanups
- remove bogus line break in Copyright header
- remove useless {} around register address definition.
- add missing whitespace before '{' in 'if' clause
Lothar Waßmann [Fri, 14 Jun 2013 13:43:58 +0000 (15:43 +0200)]
nbootce:
- don't autostart image when the load did not complete successfully
- use loff_t for flash offset
- don't use nand_read_skip_bad() since it only works when reading a
whole flash area in one go.
CONFIG_ARM_DCC_MULTI should be also removed in the patch
"serial: Remove CONFIG_SERIAL_MULTI from serial drivers"
(sha1: a3827250606895ec2dd4b8d867342b7cabf3692f)
Because the driver defines serial_* functions
which cause conflict with serial.c (multiple definition of serial_*)
Removing CONFIG_SERIAL_MULTI function also require to define
default_serial_console for cases where another serial driver
is not available in the system.
Signed-off-by: Michal Simek <michal.simek@xilinx.com> Acked-by: Marek Vasut <marex@denx.de>
Jim Lin [Thu, 24 Jan 2013 01:05:55 +0000 (01:05 +0000)]
console: USB: KBD: Fix incorrect autoboot timeout
Autoboot timeout defined by CONFIG_BOOTDELAY will not be accurate if
CONFIG_USB_KEYBOARD and CONFIG_SYS_USB_EVENT_POLL are defined in
configuration file and when tstc() function for checking key pressed
takes longer time than 10 ms (e.g., 50 ms) to finish.
Gabor Juhos [Sun, 6 Jan 2013 23:04:25 +0000 (23:04 +0000)]
common/cmd_bootm.c: prevent running of subcommands before 'bootm start'
The execution order of the bootm subcommands is fixed.
Although here is a sanity check in the state machine
which should prevent running the subcommands in wrong
order but it does not catch all possible errors.
It is possible to run any subcommand without running
'bootm start' first which leads to unexpected behaviour.
For example, running 'bootm loados' without 'bootm start'
causes a hang:
U-Boot> bootm loados
XIP Invalid Image ... OK
OK
Add a sanity check to 'do_bootm_subcommand' in order
to ensure that no subcommands can be executed before
'bootm start'.
After the patch running of any subcommand without running
'bootm start' will cause an error like this:
U-Boot> bootm loados
Trying to execute a command out of order
bootm - boot application image from memory
Alexey Brodkin [Thu, 3 Jan 2013 01:02:46 +0000 (01:02 +0000)]
drivers/block/systemace - fixed data type in "systemace_read" to match prototype in "block_dev_desc_t"
Currently we have "unsigned long blkcnt" which is fine with
CONFIG_SYS_64BIT_LBA undefined because "lbaint_t" is basically the same
"unsigned long".
If CONFIG_SYS_64BIT_LBA gets defined "lbaint_t" is defined as "unsigned
long long".
Even though not many embedded systems have CONFIG_SYS_64BIT_LBA defined
it's good to have types in function implementation that match exactly
with prototypes.
Richard Genoud [Thu, 13 Dec 2012 03:30:10 +0000 (03:30 +0000)]
FAT: remove ifdefs to make the code more readable
ifdefs in the code are making it harder to read.
The use of simple if(vfat_enabled) makes no more code and is cleaner.
(the code is discarded by the compiler instead of the preprocessor.)
NB: if -O0 is used, the code won't be discarded
and bonus, now the code compiles even if CONFIG_SUPPORT_VFAT is not
defined.
Signed-off-by: Richard Genoud <richard.genoud@gmail.com>