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>
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>
Use the MMU hardware to set up 1:1 mappings between physical and virtual
addresses. This allows us to bypass the cache when accessing the flash
without having to do any physical-to-virtual address mapping in the CFI
driver.
The virtual memory mappings are defined at compile time through a sorted
array of virtual memory range objects. When a TLB miss exception
happens, the exception handler does a binary search through the array
until it finds a matching entry and loads it into the TLB. The u-boot
image itself is covered by a fixed TLB entry which is never replaced.
This makes the 'saveenv' command work again on ATNGW100 and other boards
using the CFI driver, hopefully without breaking any rules.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
The paging system which is required to set up caching properties has not
yet been initialized when the SDRAM is initialized. So when the
map_physmem() function is converted to return the physical address
unchanged, the SDRAM initialization will break on some boards.
The avr32-specific uncached() macro will return an address which will
always cause uncached accessed to be made. Since this happens in the
board code, using avr32-specific features should be ok, and will allow
the SDRAM initialization to keep working.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
When building some avr32 boards out of tree (e.g. O=..) the linker script could
not be found. This patch references the linker script in source tree.
Signed-off-by: Andreas Biemann <biessmann@corscience.de>
A recent gcc added a new unaligned rodata section called '.rodata.str1.1',
which needs to be added the the linker script. Instead of just adding this
one section, we use a wildcard ".rodata*" to get all rodata linker section
gcc has now and might add in the future.
However, '*(.rodata*)' by itself will result in sub-optimal section
ordering. The sections will be sorted by object file, which causes extra
padding between the unaligned rodata.str.1.1 of one object file and the
aligned rodata of the next object file. This is easy to fix by using the
SORT_BY_ALIGNMENT command.
This patch has not be tested one most of the boards modified. Some boards
have a linker script that looks something like this:
*(.text)
. = ALIGN(16);
*(.rodata)
*(.rodata.str1.4)
*(.eh_frame)
I change this to:
*(.text)
. = ALIGN(16);
*(.eh_frame)
*(SORT_BY_ALIGNMENT(SORT_BY_NAME(.rodata*)))
This means the start of rodata will no longer be 16 bytes aligned.
However, the boundary between text and rodata/eh_frame is still aligned to
16 bytes, which is what I think the real purpose of the ALIGN call is.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
The AT32UC3A series of processors doesn't contain any cache, and issuing
cache control instructions on those will cause an exception. This commit
makes cacheflush.h arch-dependent in preparation for the AT32UC3A-support.
Signed-off-by: Gunnar Rangoy <gunnar@rangoy.com>
Signed-off-by: Paul Driveklepp <pauldriveklepp@gmail.com>
Signed-off-by: Olav Morken <olavmrk@gmail.com>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
This addresses all drivers whose initializers have already
been moved to board_eth_init()/cpu_eth_init().
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Replace the avr32-specific board_init_info hook by the standard
board_early_init_r hook and make it optional.
board_early_init_r() runs somewhat earlier than board_init_info used to
do, but this isn't a problem for any of the in-tree boards.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
- Separate the portmux configuration functionality from the GPIO pin
control API.
- Separate the controller-specific code from the chip-specific code.
- Allow "ganged" port configuration (multiple pins at once).
- Add more flexibility to the "canned" peripheral select functions:
- Allow using more than 23 address bits, more chip selects, as
well as NAND- and CF-specific pins.
- Make the MACB SPEED pin optional, and choose between MII/RMII
using a parameter instead of an #ifdef.
- Make it possible to use other MMC slots than slot 0, and support
different MMC/SDCard data bus widths.
- Use more reasonable pull-up defaults; floating pins may consume a
lot of power.
- Get rid of some custom portmux code from the mimc200 board code. The
old gpio/portmux API couldn't really handle its requirements, but
the new one can.
- Add documentation.
The end result is slightly smaller code for all boards. Which isn't
really the point, but at least it isn't any larger.
This has been verified on ATSTK1002 and ATNGW100. I'd appreciate if
the board maintainers could help me test this on their boards. In
particular, the mimc200 port has lost a lot of code, so I'm hoping Mark
can help me out.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Cc: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
Cc: Mark Jackson <mpfj@mimc.co.uk>
Cc: Alex Raimondi <alex.raimondi@miromico.ch>
Cc: Julien May <julien.may@miromico.ch>
Changes since v1:
* Enable pullup on NWAIT
* Add missing include to portmux-pio.h
* Rename CONFIG_PIO2 -> CONFIG_PORTMUX_PIO to match docs
Renamed initialization functions for atngw100 and atstk1000.
Removed initializations for these boards from net/eth.c
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
This patch changes the return type of initdram() from long int to phys_size_t.
This is required for a couple of reasons: long int limits the amount of dram
to 2GB, and u-boot in general is moving over to phys_size_t to represent the
size of physical memory. phys_size_t is defined as an unsigned long on almost
all current platforms.
This patch *only* changes the return type of the initdram function (in
include/common.h, as well as in each board's implementation of initdram). It
does not actually modify the code inside the function on any of the platforms;
platforms which wish to support more than 2GB of DRAM will need to modify
their initdram() function code.
Build tested with MAKEALL for ppc, arm, mips, mips-el. Booted on powerpc
MPC8641HPCN.
Signed-off-by: Becky Bruce <becky.bruce@freescale.com>
This cleans up the SDRAM initialization and related code a bit, and
allows faster booting.
* Add definitions for EBI and internal SRAM to asm/arch/memory-map.h
* Remove memory test from sdram_init() and make caller responsible
for verifying the SDRAM and determining its size.
* Remove base_address member from struct sdram_config (was sdram_info)
* Add data_bits member to struct sdram_config and kill CFG_SDRAM_16BIT
* Add support for a common STK1000 hack: 16MB SDRAM instead of 8.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Since the reset vector is always aligned to a very large boundary, we
can save a couple of KB worth of alignment padding by placing the
exception vectors at the same address.
Deciding which one it is is easy: If we're handling an exception, the
CPU is in Exception mode. If we're starting up after reset, the CPU is
in Supervisor mode. So this adds a very minimal overhead to the reset
path (only executed once) and the exception handling path (normally
never executed at all.)
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Rework the HMATRIX configuration interface so that it becomes easier
to configure the HMATRIX for boards with special needs, and add new
parts.
The HMATRIX header file has been split into a general,
chip-independent part with register definitions, etc. and a
chip-specific part with SFR bitfield definitions and master/slave
identifiers.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
This is a replacement for ATSTK1002 with 64MB SDRAM and NAND flash on
board. It's currently in production and will be available soon.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
The .flashprog section was only needed back when we were running
directly from flash, and it's even more useless on NGW100 since it
uses the CFI flash driver which never used this workaround in the
first place.
Remove it on STK1000 as well, and get rid of all the associated code and
annotations.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
The existing code assumes the SDRAM row refresh period should always
be 15.6 us. This is not always true, and indeed on the ATNGW100, the
refresh rate should really be 7.81 us.
Add a refresh_period member to struct sdram_info and initialize it
properly for both ATSTK1000 and ATNGW100. Out-of-tree boards will
panic() until the refresh_period member is updated properly.
Big thanks to Gerhard Berghofer for pointing out this issue.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
The (now obsolete) atngw100 flash programming code was having problems
programming the onboard at49bv642 chip. The atstk1000 flash
programming code may have the same bug, so import fix for this problem
from the AVR32 Linux BSP.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
With recent toolchain versions, some boards would not build because
or errors like this one (here for ocotea board when building with
ELDK 4.2 beta):
ppc_4xx-ld: section .bootpg [fffff000 -> fffff23b] overlaps section .bss [fffee900 -> fffff8ab]
For many boards, the .bss section is big enough that it wraps around
at the end of the address space (0xFFFFFFFF), so the problem will not
be visible unless you use a 64 bit tool chain for development. On
some boards however, changes to the code size (due to different
optimizations) we bail out with section overlaps like above.
The fix is to add the NOLOAD attribute to the .bss and .sbss
sections, telling the linker that .bss does not consume any space in
the image.
Signed-off-by: Wolfgang Denk <wd@denx.de>
The ATSTK1000-specific flash driver intializes bi_flashstart,
bi_flashsize and bi_flashoffset, but other flash drivers, like the CFI
driver, don't.
Initialize these in board_init_r instead so that things will still be
set up correctly when we switch to the CFI driver.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
This is a compatibility step that allows both the older form
and the new form to co-exist for a while until the older can
be removed entirely.
All transformations are of the form:
Before:
#if (CONFIG_COMMANDS & CFG_CMD_AUTOSCRIPT)
After:
#if (CONFIG_COMMANDS & CFG_CMD_AUTOSCRIPT) || defined(CONFIG_CMD_AUTOSCRIPT)
Signed-off-by: Jon Loeliger <jdl@freescale.com>
Set up the portmux for the MMC interface and enable the MMC driver
along with support for DOS partitions, ext2 and FAT filesystems.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Implement MACB initialization for AVR32 and ATSTK1000, and turn
everything on, including the MACB driver.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Conform to the "standard" interface and use initdram() instead of
board_init_memories() on AVR32. This enables us to get rid of the
sdram_size member of the global_data struct as well.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Relocate the u-boot image into SDRAM like everyone else does. This
means that we can handle much larger .data and .bss than we used to.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Rewrite the resource management code (i.e. I/O memory, clock gating,
gpio) so it doesn't depend on any global state. This is necessary
because this code is heavily used before relocation to RAM, so we
can't write to any global variables.
As an added bonus, this makes u-boot's memory footprint a bit smaller,
although some functionality has been left out; all clocks are enabled
all the time, and there's no checking for gpio line conflicts.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Add atstk1002_config target to Makefile and move the AVR32 section
down below Blackfin so that it doesn't end up in the middle of
MIPS.
Drop the autogenerated linker script thing for now. Will have to
revisit how to handle chips with different flash and RAM layout
later.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Sebastien Carlier