All architectures but MIPS are using --gc-sections on final linking.
This patch introduces that feature for MIPS to reduce the memory and
flash footprint.
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Stefan Roese <sr@denx.de>
Cc: Thomas Lange <thomas@corelatus.se>
Cc: Vlad Lungu <vlad.lungu@windriver.com>
Signed-off-by: Shinya Kuribayashi <skuribay@pobox.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>
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>
From the document, if set all arguments in "OUTPUT_FORMAT" to
"tradbigmips", then even add "-EL" to gcc we still get EB format.
pb1x00 is only used in Little-endian, so its default endian should be
set to LE.
Signed-off-by: Xiangfu Liu <xiangfu@openmobilefree.net>
Signed-off-by: Shinya Kuribayashi <skuribay@pobox.com>
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>
Most of the bss initialization loop increments 4 bytes
at a time. And the loop end is checked for an 'equal'
condition. Make the bss end address aligned by 4, so
that the loop will end as expected.
Signed-off-by: Selvamuthukumar <selva.muthukumar@e-coninfotech.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>
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>
Current assembler codes are inconsistent in the way of register jump
instruction usage; some use jr, some use j. Of course GNU as allows both
usages, but as can be expected from `Jump Register' the mnemonic `jr' is
more intuitive than `j'. For example, Linux doesn't have `j <reg>' usage
at all.
Signed-off-by: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
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>
lowlevel_init.S: Assembler messages:
lowlevel_init.S:413: Warning: Pretending global symbol used as branch target is local.
Looking at codes, the `memtest' and `clearmem' are intentional mixed
use of `global symbols' and `label' for debugging purpose. To make it
build, just disable global-symbols-use for now. As a result `memtest'
still remains as unused, but leave it be...
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
gth2.c: In function 'misc_init_r':
gth2.c:434: warning: pointer targets in passing argument 2 of 'setenv' differ in signedness
Signed-off-by: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
This patch has been sent on:
- 29 Sep 2007
Although mips_io_port_base is currently a part of IDE command, it is quite
fundamental for MIPS I/O port access such as in[bwl] and out[bwl]. So move
it to MIPS general part, and introduce `set_io_port_base()' from Linux.
This patch is triggered by multiple definition of `mips_io_port_base' build
error on gth2 (and tb0229 also needs this fix.)
board/gth2/libgth2.a(gth2.o): In function `log_serial_char':
/home/skuribay/devel/u-boot.git/board/gth2/gth2.c:47: multiple definition of `mips_io_port_base'
common/libcommon.a(cmd_ide.o):/home/skuribay/devel/u-boot.git/common/cmd_ide.c:712: first defined here
make: *** [u-boot] Error 1
Signed-off-by: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Ensure that __got_start points to top of the `.got', and __got_end points
to bottom as well, so that we never fail to count num_got_entries.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Modifications are based on the linux kernel approach and
support two use cases:
1) Add O= to the make command line
'make O=/tmp/build all'
2) Set environement variable BUILD_DIR to point to the desired location
'export BUILD_DIR=/tmp/build'
'make'
The second approach can also be used with a MAKEALL script
'export BUILD_DIR=/tmp/build'
'./MAKEALL'
Command line 'O=' setting overrides BUILD_DIR environent variable.
When none of the above methods is used the local build is performed and
the object files are placed in the source directory.
The TB5200 ("Tinybox") is a small baseboard for the TQM5200 module
integrated in a little aluminium case.
Patch by Martin Krause, 8 Jun 2006
Some code cleanup
