U-Boot Makefiles contain a number of tests for compiler features etc.
which so far are executed again and again. On some architectures
(especially ARM) this results in a large number of calls to gcc.
This patch makes sure to run such tests only once, thus largely
reducing the number of "execve" system calls.
Example: number of "execve" system calls for building the "P2020DS"
(Power Architecture) and "qong" (ARM) boards, measured as:
-> strace -f -e trace=execve -o /tmp/foo ./MAKEALL <board>
-> grep execve /tmp/foo | wc -l
Before: After: Reduction:
==================================
P2020DS 20555 15205 -26%
qong 31692 14490 -54%
As a result, built times are significantly reduced, typically by
30...50%.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Andy Fleming <afleming@gmail.com>
Cc: Kumar Gala <galak@kernel.crashing.org>
Cc: Albert Aribaud <albert.aribaud@free.fr>
cc: Graeme Russ <graeme.russ@gmail.com>
cc: Mike Frysinger <vapier@gentoo.org>
Tested-by: Graeme Russ <graeme.russ@gmail.com>
Tested-by: Matthias Weisser <weisserm@arcor.de>
Tested-by: Sanjeev Premi <premi@ti.com>
Tested-by: Simon Glass <sjg@chromium.org>
Tested-by: Macpaul Lin <macpaul@gmail.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
The examples/api is not configured with USE_PRIVATE_LIBGCC. This makes
building examples/api break on certain boards that do not/cannot use the
public libgcc.
Nevertheless, this patch has to also touch the top-level Makefile to fix
this problem because the current top-level Makefile does not specify
libgcc as a prerequisite of examples/api, and explicitly builds
examples/api _before_ libgcc.
For testing this patch, I added the following to configs/seaboard.h and
ran demo.bin on a Seaboard.
+#define CONFIG_API
+#define CONFIG_SYS_MMC_MAX_DEVICE 2
+#define CONFIG_CMD_NET
+#define CONFIG_NET_MULTI
Signed-off-by: Che-Liang Chiou <clchiou@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Noticed while building all of mpc8xx. Also
constify usage string in timer.c
Warnings fixed are:
timer.c: In function 'timer':
timer.c:189: warning: format not a string literal and no format arguments
timer.c:258: warning: format not a string literal and no format arguments
atm.c: In function 'atmUnload':
atm.c:99: warning: array subscript is above array bounds
atm.c: In function 'atmLoad':
atm.c:65: warning: array subscript is above array bounds
codec.c: In function 'codsp_write_pop_int':
codec.c:678: warning: array subscript is above array bounds
codec.c: In function 'codsp_write_cop_short':
codec.c:585: warning: array subscript is above array bounds
codec.c: In function 'codsp_write_sop_int':
codec.c:512: warning: array subscript is above array bounds
Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
GNU Makefile have two flavors of variables, recursively expanded that is
defined by using '=', and simply expanded that is defined by using ':='.
The bug is caused by using recursively expanded flavor for BIN and SREC.
As you can see below, they are prepended by $(obj) twice.
We can reproduce this bug with a simplified version of this Makefile:
$ cat >Makefile <<\EOF
obj := /path/to/obj/
ELF := hello_world
BIN_rec = $(addsuffix .bin,$(ELF)) # recursively expanded
BIN_sim := $(addsuffix .bin,$(ELF)) # simply expanded
ELF := $(addprefix $(obj),$(ELF))
BIN_rec := $(addprefix $(obj),$(BIN_rec))
BIN_sim := $(addprefix $(obj),$(BIN_sim))
show:
@echo BIN_rec=$(BIN_rec)
@echo BIN_sim=$(BIN_sim)
.PHONY: show
EOF
$ make show
BIN_rec=/path/to/obj//path/to/obj/hello_world.bin
BIN_sim=/path/to/obj/hello_world.bin
Signed-off-by: Che-Liang Chiou <clchiou@chromium.org>
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>
Rename STANDALONE_LOAD_ADDR into CONFIG_STANDALONE_LOAD_ADDR
and allow that the architecture-specific default value gets
overwritten by defining the value in the board header file.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
Cc: Daniel Hellstrom <daniel@gaisler.com>
Cc: Tsi Chung Liew <tsi-chung.liew@freescale.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Commit 44c6e65 "rename _end to __bss_end__ broke building of a large
number of systems (at least all PowerPC?):
libstubs.o: In function `app_startup':
examples/standalone/stubs.c:197: undefined reference to `__bss_end__'
The rename should not be done for the files in the
examples/standalone/ directory, as these are not using the code from
start.S, but do their own BSS clearing, and either use their own
linker scripts or the ones provided by the compilers.
Signed-off-by: Po-Yu Chuang <ratbert@faraday-tech.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>
Currently, _end is used for end of BSS section. We want _end to mean
end of u-boot image, so we rename _end to __bss_end__ first.
Signed-off-by: Po-Yu Chuang <ratbert@faraday-tech.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>
Using -fno-toplevel-reorder causes gcc to not reorder functions. This
ensures that an application's entry point will be the first function in
the application's source file.
This change, along with commit 620bbba524
should cause a standalone application's entry point to be at the base of
the compiled binary. Previously, the entry point could change depending
on gcc version and flags.
Note -fno-toplevel-reorder is only available in gcc version 4.2 or
greater.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
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>
The smc911x_detect function in /net/driver/net/smc911x.c
returns a 0 if everything was ok (a chip was found) and -1 else.
In the standalone example 'smc911x_eeprom' the return value
of smc911x_detect is interpreted in a different way
(0 for error, !0 as OK).
This leads to the error that the chip will not be detected.
Signed-off-by: Juergen Kilb <j.kilb@phytec.de>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
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>
Change the return type of the *printf() functions to the standard
"int"; no changes are needed but returning the already available
length count.
This will save a few additional strlen() calls later...
Signed-off-by: Wolfgang Denk <wd@denx.de>
Previously, standalone applications were compiled with gcc flags that
produced relocatable executables on the PowerPC architecture (eg with
the -mrelocatable and -fPIC flags). There's no reason for these
applications to be fully relocatable at this time since no relocation
fixups are performed on standalone applications.
Additionally, removing the gcc relocation flags results in the entry
point of applications residing at the base of the image. When
a standalone application was relocatable, the entry point was generally
located at an offset into the image which was confusing and prone to
errors.
This change moves the entry point of PowerPC standalone applications
from 0x40004 (usually) to 0x40000.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>
The "-ffixed-r15" option doesn't work well for gcc4. Since we
don't use gp for small data with option "-G0", we can use gp
as global data pointer. This allows compiler to use r15. It
is necessary for gcc4 to work properly.
Signed-off-by: Thomas Chou <thomas@wytron.com.tw>
Signed-off-by: Scott McNutt <smcnutt@psyent.com>
As discussed on the list, move "arch/ppc" to "arch/powerpc" to
better match the Linux directory structure.
Please note that this patch also changes the "ppc" target in
MAKEALL to "powerpc" to match this new infrastructure. But "ppc"
is kept as an alias for now, to not break compatibility with
scripts using this name.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Wolfgang Denk <wd@denx.de>
Acked-by: Detlev Zundel <dzu@denx.de>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
Cc: Peter Tyser <ptyser@xes-inc.com>
Cc: Anatolij Gustschin <agust@denx.de>
Now that the other architecture-specific lib directories have been
moved out of the top-level directory there's not much reason to have the
'_generic' suffix on the common lib directory.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Also move lib_$ARCH/config.mk to arch/$ARCH/config.mk
This change is intended to clean up the top-level directory structure
and more closely mimic Linux's directory organization.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Building examples/standalone/eepro100_eeprom triggers this error:
In file included from include/common.h:629,
from eepro100_eeprom.c:24:
include/net.h: In function 'NetReadIP':
include/net.h:430: warning: implicit declaration of function 'memcpy'
eepro100_eeprom.c: At top level:
eepro100_eeprom.c:81: error: conflicting types for 'memcpy'
include/net.h:430: error: previous implicit declaration of 'memcpy' was here
Fix this.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Building for a bf533-stamp ends up with this error:
smc91111_eeprom.o: In function `smc91111_eeprom':
examples/standalone/smc91111_eeprom.c:58: undefined reference to `memset'
make[2]: *** [smc91111_eeprom] Error 1
The new eth_struct definition means gcc has to zero out the structure on
the stack, and some gcc versions optimize this with an implicit call to
memset. So tweak the structure style to avoid that gcc feature.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Commit 6a45e38495 (Make getenv_IPaddr() global)
inadvertently added ' #include "net.h" ' to the standalone programs, creating
duplicate definitions of 'struct eth_device'. This patch removes the local
definitions and removes other code that breaks due to the change in definition.
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
According to the PPC reference implementation the udelay() function is
responsible for resetting the watchdog timer as frequently as needed.
Most other architectures do not meet that requirement, so long-running
operations might result in a watchdog reset.
This patch adds a generic udelay() function which takes care of
resetting the watchdog before calling an architecture-specific
__udelay().
Signed-off-by: Ingo van Lil <inguin@gmx.de>
Since the Makefile now controls the compilation of this, there is no need
for CONFIG checking nor the stub function.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Some versions of 'make' do not handle trailing white-spaces
properly. Trailing spaces in ELF causes a 'fake' source to
be added to the variable COBJS; leading to build failure
(listed below). The problem was found with GNU Make 3.80.
Using text-function 'strip' as a workaround for the problem.
make[1]: Entering directory `/home/sanjeev/u-boot/examples/standalone'
arm-none-linux-gnueabi-gcc -g -Os -fno-common -ffixed-r8 -msoft-float
-D__KERNEL__ -DTEXT_BASE=0x80e80000 -I/home/sanjeev/u-boot/include
-fno-builtin -ffreestanding -nostdinc -isystem /opt/codesourcery/2009q1-
203/bin/../lib/gcc/arm-none-linux-gnueabi/4.3.3/include -pipe -DCONFIG_
ARM -D__ARM__ -marm -mabi=aapcs-linux -mno-thumb-interwork -march=armv5
-Wall -Wstrict-prototypes -fno-stack-protector -g -Os -fno-common -ff
ixed-r8 -msoft-float -D__KERNEL__ -DTEXT_BASE=0x80e80000 -I/home/sanje
ev/u-boot/include -fno-builtin -ffreestanding -nostdinc -isystem /opt/co
desourcery/2009q1-203/bin/../lib/gcc/arm-none-linux-gnueabi/4.3.3/includ
e -pipe -DCONFIG_ARM -D__ARM__ -marm -mabi=aapcs-linux -mno-thumb-inte
rwork -march=armv5 -I.. -Bstatic -T u-boot.lds -Ttext 0x80e80000 -o .c
arm-none-linux-gnueabi-gcc: no input files
make[1]: *** [.c] Error 1
make[1]: Leaving directory `/home/sanjeev/u-boot/examples/standalone'
make: *** [examples/standalone] Error 2
premi #
Signed-off-by: Sanjeev Premi <premi@ti.com>
Fixed typo (s/ElF/ELF/).
Signed-off-by: Wolfgang Denk <wd@denx.de>
When the smc911x driver was converted to NET_MULTI, the smc911x eeprom was
missed. The config option needed updating as well as overhauling of the
rergister read/write functions.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Tested-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
All in-tree boards that use this controller have CONFIG_NET_MULTI
added
Also:
- changed CONFIG_DRIVER_SMC91111 to CONFIG_SMC91111
- cleaned up line lengths
- modified all boards that override weak function in this driver
- modified all eeprom standalone apps to work with new driver
- updated blackfin standalone EEPROM app after testing
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Clean up the arch/cpu/board/config checks as well as redundant setting of
srec/bin variables by using the kbuild VAR-$(...) style.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Rather than maintain/extend the current ifeq($(ARCH)) mess that exists in
the standalone Makefile, push the setting up of LOAD_ADDR out to the arch
config.mk (and rename to STANDALONE_LOAD_ADDR in the process). This keeps
the common code clean and lets the arch do whatever crazy crap it wants in
its own area.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Atmel DataFlashes by default operate with pages that are slightly bigger
than normal binary sizes (i.e. many are 1056 byte pages rather than 1024
bytes). However, they also have a "power of 2" mode where the pages show
up with the normal binary size. The latter mode is required in order to
boot with a Blackfin processor, so many people wish to convert their
DataFlashes on their development systems to this mode. This standalone
application does just that.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Since the Blackfin ABI favors higher scratch registers by default, use the
last scratch register (P3) for global data rather than the first (P5).
This allows the compiler's register allocator to use higher number scratch
P registers, which in turn better matches the Blackfin instruction set,
which reduces the size of U-Boot by more than 1024 bytes...
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Clean up the arch/cpu/board/config checks as well as redundant setting of
srec/bin variables by using the kbuild VAR-$(...) style.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Rather than maintain/extend the current ifeq($(ARCH)) mess that exists in
the standalone Makefile, push the setting up of LOAD_ADDR out to the arch
config.mk (and rename to STANDALONE_LOAD_ADDR in the process). This keeps
the common code clean and lets the arch do whatever crazy crap it wants in
its own area.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Atmel DataFlashes by default operate with pages that are slightly bigger
than normal binary sizes (i.e. many are 1056 byte pages rather than 1024
bytes). However, they also have a "power of 2" mode where the pages show
up with the normal binary size. The latter mode is required in order to
boot with a Blackfin processor, so many people wish to convert their
DataFlashes on their development systems to this mode. This standalone
application does just that.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Since the Blackfin ABI favors higher scratch registers by default, use the
last scratch register (P3) for global data rather than the first (P5).
This allows the compiler's register allocator to use higher number scratch
P registers, which in turn better matches the Blackfin instruction set,
which reduces the size of U-Boot by more than 1024 bytes...
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The attached patch corrects an error in the examples/Makefile which
causes the applications in the examples directory to hang on OMAP3
based boards. The current Makefile sets -Ttext during linking to
0x0c100000 which is outside of addressable SDRAM memory. The script
corrects the existing ifeq...else...endif logic to look at the VENDOR
tag rather than the CPU tag.
The patch affects the following configs: omap3_beagle_config,
omap3_overo_config, omap3_evm_config, omap3_pandora_config,
omap3_zoom1_config and omap3_zoom2_config.
Signed-off-by: Michael Evans <horse_dung@hotmail.com>
Edited commit message.
Signed-off-by: Wolfgang Denk <wd@denx.de>
The current files in examples are all standalone application examples,
so put them in their own subdirectory for organizational purposes
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Rather than sticking Blackfin-specific stuff into the eeprom example, use
an indirect macro so that any board can override it with their own magic
sauce in their board config file.
Also fix some spurious semi-colons in defines while I'm at it ...
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
CC: Ben Warren <biggerbadderben@gmail.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>
Wolfgang Denk
Macpaul Lin
Che-liang Chiou
Jason Hobbs
Joakim Tjernlund
Daniel Schwierzeck
Graeme Russ
Po-Yu Chuang
Mike Frysinger
Sebastien Carlier
Peter Tyser
Xiangfu Liu
Juergen Kilb
Thomas Chou
Stefan Roese
Ben Warren
Ingo van Lil
Sanjeev Premi
Scott McNutt
Sergey Mironov
Robin Getz
Michael Evans
Trent Piepho