With the ELDK 5.1 (beta) "armv7a" toolchain I'm get the following
build failure:
$ ./MAKEALL mx51evk
...
/opt/eldk-5.1/armv7a/sysroots/arm-linux-gnueabi/usr/lib/arm-linux-
gnueabi/4.6.1/libgcc.a(bpabi.o):(.ARM.exidx+0x0): undefined
reference to `__aeabi_unwind_cpp_pr1'
make: *** [u-boot] Error 1
This patch fixes the issue similar to commit
d442b6e7ad6a86e2fd0e6297291fe8872ff26fc6 but I don't know if it's
general enough or if it does harm when using other toolchains.
Signed-off-by: Wolfgang Grandegger <wg@denx.de>
U-boot itself generally builds with -nostdinc. This is because the
bootloader needs to be completely standalone. In the sandbox arch
though, we need a little bit of code to glue the u-boot world to the
host operating system, and we need to be able to access the host
libc's headers in order to do so.
Currently, we're using -I/usr/include to workaround the global
-nostdinc, but that doesn't work for everyone and for all headers.
Instead, let's filter out -nostdinc when building the os.c code.
Without this patch, some distros hit errors such as:
---8<---
In file included from /usr/include/fcntl.h:27:0,
from os.c:22:
/usr/include/features.h:323:26: fatal error:
bits/predefs.h: No such file or directory
--->8---
Signed-off-by: Andreas Bießmann <biessmann@corscience.de>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Using mmap to allocate memory from the OS for RAM simulation we can use
u-boot own malloc implementation.
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Matthias Weisser <weisserm@arcor.de>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Christian Riesch <christian.riesch@omicron.at>
Cc: Tom Rini <trini@ti.com>
Cc: Heiko Schocher <hs@denx.de>
Cc: Wolfgang Denk <wd@denx.de>
The current way memsize.c is built just made a symlink into the directory
with SPL and then compiled it like any other file there. This was bad as that
broke the out-of-tree build.
The new way introduced in this patch uses the standard spl/Makefile methods
(CONFIG_SPL_LIBCOMMON_SUPPORT / CONFIG_SPL_LIBGENERIC_SUPPORT) to let files in
common/ be built. Because common/Makefile says memsize.c is always built (SPL
and non-SPL build), this fixes our issue with memsize.c out-of-tree build.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Detlev Zundel <dzu@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
This moves SPL to common location so it can be reused by multiple boards. Also,
this commit adjusts M28 SoM to avoid breakage due to the move.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Detlev Zundel <dzu@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
i.MX6Q is freescale quad core processors with ARM cortex_a9 complex.
This patch is to add the initial support for this processor.
Signed-off-by: Jason Liu <jason.hui@linaro.org>
Cc:Stefano Babic <sbabic@denx.de>
In order to support the coming MX6 platform and to reducde
the duplicated code, we had better move some common files
or functions to the imx-common folder for sharing.
This patch does the following:
- move speed.c file from armv7/mx5/speed.c to armv7/imx-common/speed.c
- move armv7/mx5/timer.c to armv7/imx-common/timer.c, no any new feature
added but just fix the checkpatch errors in the old file and remove
the CONFIG_SYS_MX5_CLK32 reference in the file
- create one new file cpu.c file to store the common function with i.mx5/6
Signed-off-by: Jason Liu <jason.hui@linaro.org>
Cc:Stefano Babic <sbabic@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
As per new naming convention for Samsung SoC's, all Cortex-A9 and Cortex-A15
based SoC's will be classified under the name Exynos. Cortex-A9 and Cortex-A15
based SoC's will be sub-classified as Exynos4 and Exynos5 respectively.
In order to better adapt and reuse code across various upcoming Samsung Exynos
based boards, all uses of s5pc210 prefix/suffix/directory-names are renamed in
this patch. s5pc210 is renamed as exynos4210 and S5PC210/s5pc210 suffix/prefix
are renamed as exynos4/EXYNOS4.
Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
In board_init_f() the gd->bd pointer is not valid when dram_init() is called.
This only avoids dying because DRAM is at zero on Tegra2. The common ARM
routine sets up the banks in the same way anyway, so we can just remove this
code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This message is not required, since it is followed by an 'official' U-Boot
message.
U-Boot 2011.03-00048-gd7cb0d3 (May 11 2011 - 17:17:23)
TEGRA2
Board: NVIDIA Seaboard
dynamic ram_size = 1073741824
DRAM: 1 GiB
becomes:
TEGRA2
Board: NVIDIA Seaboard
DRAM: 1 GiB
This is a separate commit since it changes behavior.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Since we have cache support built in we can remove Tegra's existing cache
initialization code amd other related dead code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Some SOCs have do not start up with their 'main' CPU. The first U-Boot
code may then be executed with a CPU which does not have a CP15, or not a
useful one.
Here we split the initialization of CP15 into a separate call, which can
be performed later if required.
Once the main CPU is running, you should call cpu_init_cp15() to perform
this init as early as possible.
Existing ARMv7 boards which define CONFIG_SKIP_LOWLEVEL_INIT should not
need to change, this CP15 init is still skipped in that case. The only
impact for these boards is that the cpu_init_cp15() will be available
even if it is never used on these boards.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
The Tegra2 boot path is more complicated than it needs to be. Since we want
to move to building most of U-Boot with ARMv7 and only a small part with
ARMv4T (for AVP) it should be as simple as possible.
This makes tegra2_start() into a simple function which either does AVP
init or A9 init depending on which core is running it. Both cores now
following the same init path, beginning at _start, and the special Tegra2
boot path code is no longer required.
Only two files need to be built for ARMv4T, and this is handled in the
Tegra2 CPU Makefile.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
We want to move away from a special Tegra2 start-up, and just use
arch_cpu_init() instead. However, if we run board_init_f() from boot
we need to build it for ARMv4T, since the Tegra's AVP start-up CPU
does not support ARMv7.
The effect of this is to do the AVP init earlier, and in
arch_cpu_init(), rather that board_early_init_f().
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Sorry if this is already fixed somewhere - I could not find it.
This fixes these warnings:
usb_ohci.c: In function 'submit_control_msg':
usb_ohci.c:1081: warning: dereferencing pointer 'data_buf.76' does break strict-aliasing rules
usb_ohci.c:1081: note: initialized from here
usb_ohci.c:1084: warning: dereferencing pointer 'data_buf.76' does break strict-aliasing rules
usb_ohci.c:1084: note: initialized from here
usb_ohci.c:1087: warning: dereferencing pointer 'data_buf.76' does break strict-aliasing rules
usb_ohci.c:1087: note: initialized from here
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Fix:
board.c: In function 'board_init_r':
board.c:262:8: warning: unused variable 's' [-Wunused-variable]
(raised for configurations without ethernet support like
vct_premium_small)
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
Fix all errors and all warnings except for "externs should be avoided"
which could require more extensive changes.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
Standard Debian powerpc and powerpcspe systems only include hard-float
libgcc in their native compilers, which causes scary build warnings when
building U-Boot.
Debian and other PowerPC-supporting distributions used to provide libgcc
and other libraries in a "nof" (soft-float) form in the "multilib"
packages. As they were completely unused by the distribution and
therefore tended to be very buggy it was decided to save some time on
the part of the maintainers and build-servers by removing them.
Admittedly, right now the linker warnings do not indicate any problems,
as the included routines do not use any floating point at all.
The concern is that if floating-point code were ever added it might
cause hard-float code to be unexpectedly included in U-Boot without
generating a hard error. This would cause unexplained crashes or
indeterminate results at runtime.
The easiest way to resolve this is to borrow the routines that U-Boot
needs from the Linux kernel, which has the same issue.
Specifically, the routines are: _ashldi3(), _ashrdi3(), and _lshrdi3().
They were borrowed from arch/powerpc/kernel/misc_32.S as of v2.6.38-rc5,
commit 85e2efbb1db9a18d218006706d6e4fbeb0216213, and are GPLv2+.
The Makefile framework was copied from the U-Boot ARM port.
Signed-off-by: Kyle Moffett <Kyle.D.Moffett@boeing.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Kim Phillips <kim.phillips@freescale.com>
Cc: Andy Fleming <afleming@gmail.com>
Cc: Kumar Gala <kumar.gala@freescale.com>
Cc: Stefan Roese <sr@denx.de>
Fix:
clocks.c: In function 'setup_post_dividers':
clocks.c:175: warning: comparison is always true due to limited range of
data type
clocks.c:177: warning: comparison is always true due to limited range of
data type
clocks.c:179: warning: comparison is always true due to limited range of
data type
clocks.c:181: warning: comparison is always true due to limited range of
data type
clocks.c:183: warning: comparison is always true due to limited range of
data type
clocks.c:185: warning: comparison is always true due to limited range of
data type
clocks.c:187: warning: comparison is always true due to limited range of
data type
clocks.c:189: warning: comparison is always true due to limited range of
data type
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: sricharan <r.sricharan@ti.com>
Cc: Tom Rini <trini@ti.com>
Fix:
clocks-common.c: In function 'setup_dplls':
clocks-common.c:256:6: warning: variable 'sysclk_ind' set but not used
[-Wunused-but-set-variable]
clocks-common.c: In function 'setup_non_essential_dplls':
clocks-common.c:292:6: warning: variable 'sysclk_ind' set but not used
[-Wunused-but-set-variable]
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: sricharan <r.sricharan@ti.com>
Cc: Tom Rini <trini@ti.com>
Acked-by: Marek Vasut <marek.vasut@gmail.com>
Acked-by: Tom Rini <trini@ti.com>
Fix:
spl.c: In function 'jump_to_image_no_args':
spl.c:103: warning: assignment makes pointer from integer without a cast
spl.c:105: warning: dereferencing type-punned pointer will break
strict-aliasing rules
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: sricharan <r.sricharan@ti.com>
Cc: Tom Rini <trini@ti.com>
Acked-by: Marek Vasut <marek.vasut@gmail.com>
Acked-by: Tom Rini <trini@ti.com>
The serial driver for iMX SOCs is continuosly changed if a
new SOC or not yet used port is used. CONFIG_SYS_<SOC>_<UART Port>
defines were used only to find the base address for the selected UART.
Instead of that, move the base address to the board configuration
file and drop all #ifdef from driver.
Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Marek Vasut <marek.vasut@gmail.com>
CC: Wolfgang Denk <wd@denx.de>
CC: Fabio Estevam <fabio.estevam@freescale.com>
CC: Helmut Raiger <helmut.raiger@hale.at>
CC: John Rigby <jcrigby@gmail.com>
CC: Matthias Weisser <weisserm@arcor.de>
CC: Jason Liu <jason.hui@linaro.org>
Acked-by: Jason Liu <jason.hui@linaro.org>
Some functions were internal to the apbh dma driver, so annotate them static.
Some of the functions weren't used at all so drop them. This makes the U-Boot
binary smaller by about 1500 bytes.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Wolfgang Denk <wd@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
This corects the warning below, obtained with my gcc 4.6 compiler.
arch/arm/cpu/armv7/mx5/libmx5.o: In function `decode_pll':
arch/arm/cpu/armv7/mx5/clock.c:94: undefined reference to `__aeabi_uldivmod'
I am not able to test this on MX5x hardware, but it does improve the
MAKEALL output for me. You may already have a similar patch, but I cannot
see it on the list.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add an SPL_BOARD_INIT hook and for OMAP3 have it turn on i2c. OMAP4
doesn't need i2c enabled in SPL. Enable SPL_BOARD_INIT on devkit8000.
Cc: Frederik Kriewitz <frederik@kriewitz.eu>
Signed-off-by: Tom Rini <trini@ti.com>
Add Hynix 200MHz timing information to <asm/arch-omap3/mem.h>.
This also changes CONFIG_SYS_TEXT_BASE to 0x80100000.
Signed-off-by: Tom Rini <trini@ti.com>
This introduces 200MHz Micron parts timing information based on x-loader
to <asm/arch-omap3/mem.h> and Numonyx MCFG calculation. The memory init
logic is also based on what x-loader does in these cases. Note that
while previously u-boot would be flashed in with SW ECC in this case it
now must be flashed with HW ECC. We also change CONFIG_SYS_TEXT_BASE to
0x80100000.
Cc: Dirk Behme <dirk.behme@gmail.com>
Beagleboard rev C5, xM rev A:
Tested-by: Tom Rini <trini@ti.com>
Beagleboard xM rev C:
Tested-by: Matt Ranostay <mranostay@gmail.com>
Beagleboard rev B7, C2, xM rev B:
Tested-by: Matt Porter <mporter@ti.com>
Signed-off-by: Tom Rini <trini@ti.com>
A number of boards are populated with a PoP chip for both DDR and NAND
memory. Other boards may simply use this as an easy way to identify
board revs. So we provide a function that can be called early to reset
the NAND chip and return the result of NAND_CMD_READID. All of this
code is put into spl_id_nand.c and controlled via CONFIG_SPL_OMAP3_ID_NAND.
Signed-off-by: Tom Rini <trini@ti.com>
This changes to making the board be responsible for providing the
memory initialization timings in SPL and converts the devkit8000
to this framework. In SPL we try and initialize both CS0 and CS1.
Cc: Frederik Kriewitz <frederik@kriewitz.eu>
Signed-off-by: Tom Rini <trini@ti.com>
This adds the optimal SDRC autorefresh control register values for
100Mhz, 133MHz, 165MHz and 200MHz clocks. We switch to using this
to provide the default 165MHz value.
Signed-off-by: Tom Rini <trini@ti.com>
This adds an MCFG macro to calculate the correct value, similar to
the ACTIMA/ACTIMB macros and adds a comment that all of the potential
values here are documented in the TRM. Then we convert the Micron
value to use this macro.
Signed-off-by: Tom Rini <trini@ti.com>
It's possible to need to call this function on the same banks multiple
times so we want to be sure that 'pos A' is cleared out again at the
end.
Signed-off-by: Tom Rini <trini@ti.com>
Since we go through the sequence to setup the SDRC timings more than
once, break this logic out into its own function and have that function
call mem_ok() to make sure the memory is usable.
Signed-off-by: Tom Rini <trini@ti.com>
We update the comment in make_cs1_contiguous() to be a little bit
more clear (it's been copy/pasted from other silicons) and then
explain in dram_init() why we need to always try this.
Note that in the previous behavior we were always calling this on
boards that never had cs1 populated anyhow so making sure we do
this always is fine and will correct things like omap3evm detecting
an invalid amount of memory (384MB).
Signed-off-by: Tom Rini <trini@ti.com>
Expand the "enable the config" comment to explain what the bit shifts
are and define out two of the magic numbers.
Signed-off-by: Tom Rini <trini@ti.com>
This patch avoids build breakage for SPLs that do not support printf.
Signed-off-by: Christian Riesch <christian.riesch@omicron.at>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Acked-by: Tom Rini <trini@ti.com>
Acked-by: Andreas Bießmann <andreas.devel@googlemail.com>
move the board/davinci/common/misc.c file to
arch/arm/cpu/arm926ejs/davinci/misc.c, so all
davinci boards can use this functions.
Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
Cc: Tom Rini <tom.rini@gmail.com>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Christian Riesch <christian.riesch@omicron.at>
move struct davinci_rtc to arch/arm/include/asm/arch-davinci/hardware.h
and add RTC_KICK0R_WE, RTC_KICK1R_WE defines,
so they are global useable.
Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
This patch replaces the pinmux configuration code in
arch/arm/cpu/arm926ejs/davinci/da850_lowlevel.c by the code from
arch/arm/cpu/arm926ejs/davinci/pinmux.c.
Signed-off-by: Christian Riesch <christian.riesch@omicron.at>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
Acked-by: Heiko Schocher <hs@denx.de>
Up to now nearly every davinci board has separate code for the
definition of pinmux configurations. This patch adds pinmux
configurations for the DA850 SoCs to the arch tree which may later
be used for all DA850 based boards.
Signed-off-by: Christian Riesch <christian.riesch@omicron.at>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
Cc: Heiko Schocher <hs@denx.de>
Cc: Mike Frysinger <vapier@gentoo.org>
Acked-by: Heiko Schocher <hs@denx.de>
always do the cpu critical inits in cpu_init_crit,
and only jump to lowlevel_init, if CONFIG_SKIP_LOWLEVEL_INIT
is not defined.
Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Wolfgang Denk <hs@denx.de>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
Cc: Tom Rini <tom.rini@gmail.com>
Cc: Christian Riesch <christian.riesch@omicron.at>
Added noop implementation for dcache operations that will buzz
about missing real implementation and disable the dcache.
This fixes compilation of DaVinci EMAC driver on arm926ejs.
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
DaVinci EMAC is found not only on DaVinci SoCs but on some OMAP3 SoCs
also. This patch moves common defines from arch-davinci/emac_defs.h to
drivers/net/davinci_emac.h
DaVinci specific PHY drivers hacked to include the new header. We might
want to switch to phylib in future.
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
The value from TRIM is not working for some 4430 silicons.
So, override with hw team recommended value. However, for
4460 TRIM value shall be used as long as the part is trimmed
This fixes boot problem on some OMAP4430 ES2.0 Panda boards
out there.
Cc: Steve Sakoman <sakoman@gmail.com>
Signed-off-by: Aneesh V <aneesh@ti.com>
There was a typo in the EMIF driver. It went un-noticed
because it affected only when automatic detection is enabled
and even then half the memory was configured and identified
properly.
Reported-by: Rockefeller <rockefeller.lin@innocomm.com>
Signed-off-by: Aneesh V <aneesh@ti.com>
TPS power IC is controlled using a GPIO (gpio_wk7).
This GPIO should be maintained at logic 1 always. As
such an internal pull-up on this pin will do the job,
driving the GPIO outuput is not needed. This will avoid
the need of using GPIO library in SPL and also may
save some power.
Signed-off-by: Aneesh V <aneesh@ti.com>
The vector is not correctly setup in armv7 except for OMAP3.
Correcting this.
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Signed-off-by: Aneesh V <aneesh@ti.com>
This allows SPL to have default implementation of
save_boot_params(), useful for SoCs that do
not intend to override this default implementation
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Signed-off-by: Aneesh V <aneesh@ti.com>
We were not disabling external caches before jumping
to kernel. We were flushing all caches including
external caches and disabling caches globally in
CP15 System Control register. Apparently this is not
enough.
The bootstrap loader in Linux kernel that does decompression
enables data-caches again, flush them after use and disable
them before jumping to kernel proper. However, it's not aware
of the external caches.
Since we have left external cache enabled, external cache will
get used once caches are enabled globally, but it's not flushed
because decompressor is not aware of external caches. When it
jumps to kernel with caches disabled globally, we have stale
data in the external cache and a coherency problem.
This was breaking the boot for OMAP4 with latest mainline
kernel. The solution is to disable external caches in
cleanup_before_linux(). With this fix kernel is booting again.
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Signed-off-by: Aneesh V <aneesh@ti.com>
This patch fixes the clear bss loop for bss sections that have
zero length, i.e., where __bss_start == __bss_end__.
Signed-off-by: Christian Riesch <christian.riesch@omicron.at>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Stefan Herbrechtsmeier <sherbrec@cit-ec.uni-bielefeld.de>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
The start.S on PXA was very obscure. This reworks it back to be close to arm1136
start.S and others.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
V2: Don't compile in relocation support if building SPL
Building for some m68k boards results in the warning:
cpu_init.c: In function 'cpu_init_f':
cpu_init.c:287: warning: suggest parentheses around
operand of '!' or change '&' to '&&' or '!' to '~'
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Recent builds for SH4 boards fail with a lot of errors like:
cmd_mem.o: In function 'dcache_invalid_range':
include/asm/cache.h:25: multiple definition of 'dcache_invalid_range'
include/asm/cache.h:25: first defined here
This is due to the funcs being defined in the header, but not static
or inline or extern. So move them to the sh4-specific cache.c file.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
The ashrsi3 function is used by some commands that aren't in SH2A
default configs (e.g. JFFS2).
The ashrsi3.S file has been copied from Linux.
Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Not all SuperH toolchains support -mno-fdpic. Chances are good that if
the flag doesn't work, it isn't defaulting to the FDPIC ABI, so the flag
isn't needed. So only add it if it is actually supported.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Fix:
cpu.c: In function 'checkcpu':
cpu.c:51:7: warning: variable 'ver' set but not used
[-Wunused-but-set-variable]
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
* 'master' of git://git.denx.de/u-boot-mpc85xx:
mpc85xx: support for Freescale COM Express P2020
arch/powerpc/cpu/mpc8xxx/ddr/interactive.c: Fix GCC 4.6 build warning
mpc85xx: support board-specific reset function
powerpc/85xx: verify the localbus device tree address before booting the OS
mpc8xxx: update module_type values from JEDEC DDR3 SPD Specification
powerpc/p3060qds: Add board related support for P3060QDS platform
powerpc/85xx: clean up and document the QE/FMAN microcode macros
powerpc/85xx: always implement the work-around for Erratum SATA_A001
powerpc/85xx: CONFIG_FSL_SATA_V2 should be defined in config_mpc85xx.h
powerpc/85xx: Add workaround for erratum A-003474
powerpc/85xx: fixup flexcan device tree clock-frequency
powerpc/85xx: Add workaround for erratum CPU-A003999
This adds support for the Freescale COM Express P2020 board. This board
is similar to the P1_P2_RDB, but has some extra (as well as missing)
peripherals.
Unlike all other mpc85xx boards, it uses a watchdog timeout to reset.
Using the HRESET_REQ register does not work.
This board has no NOR flash, and can only be booted via SD or SPI. This
procedure is documented in Freescale Document Number AN3659 "Booting
from On-Chip ROM (eSDHC or eSPI)." Some alternative documentation is
provided in Freescale Document Number P2020RM "P2020 QorIQ Integrated
Processor Reference Manual" (section 4.5).
Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Fix:
interactive.c: In function 'fsl_ddr_interactive':
interactive.c:1357:15: warning: variable 'len' set but not used [-Wunused-but-set-variable]
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
This is useful for boards which cannot be reset in the usual way for the
85xx CPU. An example is a board which can only be reset by a hardware
watchdog.
Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The localbus controller node in the device tree is typically a root node,
even though the controller is part of CCSR. If we were to put the lbc
node under the SOC node, then the 'ranges' property in the lbc node would
translate through the 'ranges' property of the parent SOC node, and we
don't want that.
Since the lbc is a separate node, it's possible for the 'reg' property to
be wrong. This happened with the original version of p1022ds.dts, which
used a 32-bit value in the 'reg' address, instead of a 36-bit address.
Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Newer JEDEC DDR3 SPD Specifications define several additional values for
the DDR3 module_type field which were undefined when this code was
written. Update the code to handle the newer module types.
Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu>
Cc: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Several macros are used to identify and locate the microcode binary image
that U-boot needs to upload to the QE or Fman. Both the QE and the Fman
use the QE Firmware binary format to package their respective microcode data,
which is why the same macros are used for both. A given SOC will only have
a QE or an Fman, so this is safe.
Unfortunately, the current macro definition and usage has inconsistencies.
For example, CONFIG_SYS_FMAN_FW_ADDR was used to define the address of Fman
firmware in NOR flash, but CONFIG_SYS_QE_FW_IN_NAND contains the address
of NAND. There's no way to know by looking at a variable how it's supposed
to be used.
In the future, the code which uploads QE firmware and Fman firmware will
be merged.
Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
On the P1022/P1013, the work-around for erratum SATA_A001 was implemented
only if U-Boot initializes SATA, but SATA is not initialized by default. So
move the work-around to the CPU initialization function, so that it's always
executed on the SOCs that need it.
Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Macro CONFIG_FSL_SATA_V2 is defined if the SOC has a V2 Freescale SATA
controller, so it should be defined in config_mpc85xx.h instead of the various
board header files. So now CONFIG_FSL_SATA_V2 is always defined on the P1013,
P1022, P2041, P3041, P5010, and P5020. It was already defined for the
P1010 and P1014.
Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Erratum A-003474: Internal DDR calibration circuit is not supported
Impact:
Experience shows no significant benefit to device operation with
auto-calibration enabled versus it disabled. To ensure consistent timing
results, Freescale recommends this feature be disabled in future customer
products. There should be no impact to parts that are already operating
in the field.
Workaround:
Prior to setting DDR_SDRAM_CFG[MEM_EN]=1, do the following:
1. Write a value of 0x0000_0015 to the register at offset
CCSRBAR + DDR OFFSET + 0xf30
2. Write a value of 0x2400_0000 to the register at offset
CCSRBAR + DDR OFFSET + 0xf54
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Make the fixup matchable with dts and kernel. Update the compatible from
"fsl,flexcan-v1.0" to "fsl,p1010-flexcan" and Change the "clock-freq"
property to "clock-frequency". We also change flexcan frequency from
CCB-clock to CCB-clock/2 according to P1010 spec.
We now keep the old interfaces to make previous kernel work. They should
be removed in the future.
Signed-off-by: Jia Hongtao <B38951@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Erratum A-003999: Running Floating Point instructions requires special
initialization.
Impact:
Floating point arithmetic operations may result in an incorrect value.
Workaround:
Perform a read modify write to set bit 7 to a 1 in SPR 977 before
executing any floating point arithmetic operation. This bit can be set
when setting MSR[FP], and can be cleared when clearing MSR[FP].
Alternatively, the bit can be set once at boot time, and never cleared.
There will be no performance degradation due to setting this bit.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
When gcc compiles some 64 bit operations on a 32 bit machine, it generates
calls to small functions instead of instructions which do the job directly.
Those functions are defined in libgcc and transparently provide whatever
functionality was necessary. Unfortunately, u-boot can be built with a
non-standard ABI when libgcc isn't. More specifically, u-boot uses
-mregparm. When the u-boot and libgcc are linked together, very confusing
bugs can crop up, for instance seemingly normal integer division or modulus
getting the wrong answer or even raising a spurious divide by zero
exception.
This change borrows (steals) a technique and some code from coreboot which
solves this problem by creating wrappers which translate the calling
convention when calling the functions in libgcc. Unfortunately that means
that these instructions which had already been turned into functions have
even more overhead, but more importantly it makes them work properly.
To find all of the functions that needed wrapping, u-boot was compiled
without linking in libgcc. All the symbols the linker complained were
undefined were presumed to be the symbols that are needed from libgcc.
These were a subset of the symbols covered by the coreboot code, so it was
used unmodified.
To prevent symbols which are provided by libgcc but not currently wrapped
(or even known about) from being silently linked against by code generated
by libgcc, a new copy of libgcc is created where all the symbols are
prefixed with __normal_. Without being purposefully wrapped, these symbols
will cause linker errors instead of silently introducing very subtle,
confusing bugs.
Another approach would be to whitelist symbols from libgcc and strip out
all the others. The problem with this approach is that it requires the
white listed symbols to be specified three times, once for objcopy, once so
the linker inserts the wrapped, and once to generate the wrapper itself,
while this implementation needs it to be listed only twice. There isn't
much tangible difference in what each approach produces, so this one was
preferred.
Signed-off-by: Gabe Black <gabeblack@chromium.org>
The new implementation is about twice as fast as the old. This is from
glibc-2.14, sysdeps/i386/memset.c.
Signed-off-by: Gabe Black <gabeblack@chromium.org>
This change adds an upper bound for symbols which are fixed up after u-boot
is relocated into RAM. This way portions that are left at their original
location can be referred to without having to manually fix up any pointers.
Signed-off-by: Gabe Black <gabeblack@chromium.org>
There are two blobs embedded into the u-boot image which are linked to run
at an address which is different from where they actually end up in the
ROM, one called "realmode" and one called "bios". There are realmode_setup
and bios_setup functions which prepare those blobs by copying them into the
location they're supposed to run from, among other things.
During u-boot relocation from ROM to RAM, the text and a few data segments
are copied over. The realmode and bios sections are not copied, and so the
only place they can be read from is their original location in the ROM.
Looking specifically at the bios blob, there are symbols defined in the
linker script called __bios_start and __bios_size which are defined to be
the start and size of the blob in the ROM.
In the bios_setup function, there seem to be two mistakes happening. First,
the offset from ROM to RAM is being added to __bios_start which implies that
this code expects to use the copy moved to RAM. No such copy is made, so
that's wrong. More subtly, when u-boot relocates itself, it goes through
all of the relocations stored in .rel.dyn and fixes them up. This has the
effect of transforming the __bios_start reference in bios_setup so that it
refers to the version in RAM (if one existed) instead of the one in ROM. To
correct for that, the offset actually needs to be subtracted out again to
translate the address back into the ROM.
The net effect is that for both blobs, a + needs to be changed to a -.
Signed-off-by: Gabe Black <gabeblack@chromium.org>
