r9 is a platform-specific register in ARM EABI and not per
definition a general purpose register. Do not use it while
relocating so it can be used for gd.
cc: Albert ARIBAUD <albert.u.boot@aribaud.net> Signed-off-by: Jeroen Hofstee <jeroen@myspectrum.nl>
Michal Simek [Wed, 28 Aug 2013 05:36:31 +0000 (07:36 +0200)]
arm: zynq: Fix timer loadaddress
Reload address was written to the counter register
instead of load register.
The problem happens when timer expires but never
reload to ~0UL (it is downcount timer).
Reported-by: Stephen MacMahon <stephenm@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Jeroen Hofstee [Sat, 24 Aug 2013 11:55:38 +0000 (13:55 +0200)]
arm: prevent using movt/movw address loads
The movt/movw instruction can be used to hardcode an
memory location in the instruction itself. The linker
starts complaining about this if the compiler decides
to do so: "relocation R_ARM_MOVW_ABS_NC against `a local
symbol' can not be used" and it is not support by U-boot
as well. Prevent their use by requiring word relocations.
This allows u-boot to be build at other optimalization
levels then -Os.
Signed-off-by: Jeroen Hofstee <jeroen@myspectrum.nl> Cc: TigerLiu@viatech.com.cn Cc: Albert ARIBAUD <albert.u.boot@aribaud.net> Acked-by: Simon Glass <sjg@chromium.org>
am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot.
If, in CONFIG_BOOTCOMMAND, the environment switches both the mmcdev
and bootpart variables to refer to MMC device 1, it would make sense
that the mmcroot env variable should switch to that device as well.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
if phy_connect() did not find a phy, phydev is NULL and
following code in cpsw_phy_init() crashes. Fix this.
Signed-off-by: Heiko Schocher <hs@denx.de> Cc: Joe Hershberger <joe.hershberger@gmail.com> Cc: Mugunthan V N <mugunthanvnm@ti.com> Cc: Tom Rini <trini@ti.com> Acked-by: Mugunthan V N <mugunthanvnm@ti.com>
Lokesh Vutla [Fri, 23 Aug 2013 12:04:17 +0000 (17:34 +0530)]
ARM: OMAP5: Avoid writing into LDO SRAM bits
Writing magic bits into LDO SRAM was suggested only for OMAP5432
ES1.0. Now these are no longer applicable. Moreover these bits should
not be overwritten as they are loaded from EFUSE. So avoid
writing into these registers.
Steve Kipisz [Wed, 14 Aug 2013 14:51:31 +0000 (10:51 -0400)]
am335x:Handle worst case scenario for Errata 1.0.24
In Errata 1.0.24, if the board is running at OPP50 and has a warm reset,
the boot ROM sets the frequencies for OPP100. This patch attempts to
drop the frequencies back to OPP50 as soon as possible in the SPL. Then
later the voltages and frequencies up set higher.
Cc: Enric Balletbo i Serra <eballetbo@iseebcn.com> Cc: Lars Poeschel <poeschel@lemonage.de> Signed-off-by: Steve Kipisz <s-kipisz2@ti.com>
[trini: Adapt to current framework] Signed-off-by: Tom Rini <trini@ti.com>
Tom Rini [Fri, 30 Aug 2013 20:28:46 +0000 (16:28 -0400)]
am335x_evm: am33xx_spl_board_init function and scale core frequency
Add a am33xx_spl_board_init (and enable the PMICs) that we may see,
depending on the board we are running on. In all cases, we see if we
can rely on the efuse_sma register to tell us the maximum speed. In the
case of Beaglebone White, we need to make sure we are on AC power, and
are on later than rev A1, and then we can ramp up to the PG1.0 maximum
of 720Mhz. In the case of Beaglebone Black, we are either on PG2.0 that
supports 1GHz or PG2.1. As PG2.0 may or may not have efuse_sma set, we
cannot rely on this probe. In the case of the GP EVM, EVM SK and IDK we
need to rely on the efuse_sma if we are on PG2.1, and the defaults for
PG1.0/2.0.
Fix wrong sdhci host control register read and write
The patch fixes the improper read and write of sdhci
host control register for sdma transfer.
The problem comes when reading and writing 1 byte long
host control register with the sdhci_readl() and
sdhci_writel(). The misuse of these functions overwrite
the value of the next registers which are in 4 bytes boundary.
This patch replaces four byte register read/write functions
with one byte read/write ones. Beside, it eliminates
unnecessary bit operation. i.e. or-ing zero against a variable.
Without those it's very easy to make mistakes when for instance
the 'size' field is more than just a constant.
Signed-off-by: Mischa Jonker <mjonker@synopsys.com> Cc: Alexey Brodkin <abrodkin@synopsys.com> Cc: Marek Vasut <marex@denx.de> Cc: Anton Staaf <robotboy@chromium.org> Cc: Tom Rini <trini@ti.com> Cc: Wolfgang Denk <wd@denx.de>
mmc/dw_mmc: Allocate the correct amount of descriptors
This fixes two issues:
* a descriptor was allocated for every block, while a descriptor can
take 8 blocks
* there was an off-by-one error in the descriptor preparation: there
were two last descriptors, one with length==0
In dwmci_prepare_data, the descriptors are allocated for DMA transfer.
These are allocated using the ALLOC_CACHE_ALIGN_BUFFER. This macro uses
the stack to allocate these descriptors. This becomes a problem if the
DMA transfer continues after the processor leaves the function in which
the descriptors were allocated.
Therefore, I have moved the allocated of the buffers up one level, to
dwmci_send_cmd(). The DMA transfer should be complete when leaving this
function.
Paul Burton [Mon, 9 Sep 2013 14:30:26 +0000 (15:30 +0100)]
mmc: don't support write & erase for SPL builds
For SPL builds this is just dead code since we'll only need to read.
Eliminating it results in a significant size reduction for the SPL
binary, which may be critical for certain platforms where the binary
size is highly constrained.
Signed-off-by: Paul Burton <paul.burton@imgtec.com> Acked-by: Pantelis Antoniou <panto@antoniou-consulting.com>
doc: README.mxs: Add instruction to install 'libssl-dev'
Since commit bce883707 (ARM: mxs: tools: Add mkimage support for MXS bootstream)
the following build error is seen when doing a MAKEALL build:
$ ./MAKEALL mx28evk
Configuring for mx28evk - Board: mx28evk, Options: ENV_IS_IN_MMC
mxsimage.c:18:25: fatal error: openssl/evp.h: No such file or directory
Add an entry about the need of installing the 'libssl-dev' package.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Acked-by: Marek Vasut <marex@denx.de>
mx6sabresd: Avoid hang when HDMI cable is not connected
Since commit d9b894603 (mx6sabresd: Add LVDS splash screen support) the
following hang happens if the HDMI cable is not connected or the 'panel'
variable is not set:
wandboard: Use imx6dl-wandboard.dtb for the solo version
The wandboard solo version should boot the 'imx6dl-wandboard.dtb' file, since
dual-lite and solo variants are the same SoC with only the number of cores being
different.
Tom Rini [Fri, 30 Aug 2013 20:28:45 +0000 (16:28 -0400)]
am33xx: Add the efuse_sma CONTROL_MODULE register
Starting with PG2.1 we have a register in the CONTROL_MODULE that is set
with the package type and maximum supported frequency. Add this, and
the relevant mask/values.
Tom Rini [Fri, 30 Aug 2013 20:28:44 +0000 (16:28 -0400)]
am33xx: Add am33xx_spl_board_init function, call
We need to allow for a further call-out in spl_board_init. Call this
am33xx_spl_board_init and add a __weak version. This function may be
used to scale the MPU frequency up, depending on board needs.
Philip, Avinash [Fri, 30 Aug 2013 20:28:43 +0000 (16:28 -0400)]
drivers/power/pmic: Add tps65910 driver
Add a driver for the TPS65910 PMIC that is found in the AM335x GP EVM,
AM335x EVM SK and others.
Signed-off-by: Philip, Avinash <avinashphilip@ti.com>
[trini: Split and rework Avinash's changes into new drivers/power
framework] Signed-off-by: Tom Rini <trini@ti.com>
Greg Guyotte [Fri, 30 Aug 2013 20:28:42 +0000 (16:28 -0400)]
drivers/power/pmic: Add tps65217 driver
Add a driver for the TPS65217 PMIC that is found in the Beaglebone
family of boards.
Signed-off-by: Greg Guyotte <gguyotte@ti.com>
[trini: Split and rework Greg's changes into new drivers/power
framework] Signed-off-by: Tom Rini <trini@ti.com>
Wolfgang Denk [Tue, 17 Sep 2013 09:24:06 +0000 (11:24 +0200)]
SPDX: fix IBM-pibs license identifier
The SPDX License List version 1.19 now contains an official entry for
the IBM-pibs license. However, instead of our suggestion "ibm-pibs",
the SPDX License List uses "IBM-pibs", with the following rationale:
"The reason being that all other SPDX License List short identifiers
tend towards using capital letters unless spelling a word. I'd prefer
to be consistent to this end".
Mark Langsdorf [Tue, 10 Sep 2013 20:14:56 +0000 (15:14 -0500)]
part_efi: make sure the gpt_pte is freed
the gpt_pte wasn't being freed if it was checked against an invalid
partition. The resulting memory leakage could make it impossible
to repeatedly attempt to load non-existent files in a script.
Also, downgrade the message for not finding an invalid partition
from a printf() to a debug() so as to minimize message spam in
perfectly normal situations.
Signed-off-by: Mark Langsdorf <mark.langsdorf@calxeda.com>
Netconsole calls eth_halt() before giving control to another operating
system.
But the state machine of netconsole don't take it into account.
Thus, netconsole calls network functions of an halted network device,
making the whole system freeze.
Rather than modifying the state machine of netconsole, we just unregister
the current network device before booting. It does work because
nc_send_packet() verifies that the current network device is not null.
Paul Burton [Fri, 6 Sep 2013 10:23:55 +0000 (11:23 +0100)]
bootm: use BOOTM_STATE_OS_CMDLINE flag for plain bootm
A plain bootm used to call the architecture specific boot function with
no flags, but was modified by commit 35fc84fa "Refactor the bootm
command to reduce code duplication" to call the architecture specific
boot function multiple times with various flags in sequence. The
BOOTM_STATE_OS_CMDLINE flag was not used, indeed it seems that at least
ARM prepares the command line on BOOTM_STATE_OS_PREP. However on MIPS
since commit 59e8cbdb "MIPS: bootm: refactor initialisation of kernel
cmdline" the command line is not prepared in response to a
BOOTM_STATE_OS_PREP flag, only on BOOTM_STATE_OS_CMDLINE or a call with
no flags. The end result is that a combination of those 2 commits leads
to MIPS boards booting kernels with no command line arguments.
An extra invocation of the architecture specific boot function with
BOOTM_STATE_OS_CMDLINE fixes this.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
arm: atmel: cpux9k2: increase malloc space to fix crash on start u-boot
Since UBIFS is enabled for cpux9k2, more malloc space is needed.
For the current uboot 2013.10-rcX the size is to small, this will fix the
startup problems by increasing the malloc space to 4MiB.
Signed-off-by: Jens Scharsig (BuS Elektronik) <esw@bus-elektronik.de> Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
Paul Burton [Wed, 4 Sep 2013 15:12:26 +0000 (16:12 +0100)]
mmc: size optimization when !CONFIG_MMC_SPI
When CONFIG_MMC_SPI is not enabled, the MMC_MODE_SPI capability can
never be set. However there is code in mmc.c which uses the
mmc_host_is_spi macro to check that capability & act accordingly. If we
expand that macro to 0 when CONFIG_MMC_SPI is not set (since it will
always be 0 at runtime anyway) then the compiler can optimize away the
SPI-specific code paths in mmc.c.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Lubomir Popov [Wed, 14 Aug 2013 15:59:18 +0000 (18:59 +0300)]
ARM: OMAP: Enable 8-bit eMMC access for OMAP4/5/DRA7xx
Enable 8-bit host capability for HSMMC2 and/or HSMMC3. CONFIG_HSMMC2_8BIT
(for OMAP4/5/DRA7xx) and/or CONFIG_HSMMC3_8BIT (for DRA7xx only) must be
defined in the board header if an 8-bit eMMC device is connected to the
corresponding port.
Fix the "No status update" error that appeared for eMMC devices by
inserting a 20 us delay between writing arguments and command. This
solution has been proposed by Michael Cashwell <mboards@prograde.net>.
"mmc_send_cmd: timeout: No status update" error sometimes happens in
omap_hsmmc driver func mmc_send_cmd() when the MMC controller card
identification and selection sequence is executed for eMMC on OMAP4
boards.
It happens due to incorrect execution of CMD line reset procedure
for OMAP4. Because CMD(DAT) lines reset procedures are slightly
different for OMAP3 and OMAP4(AM335x,OMAP5,DRA7xx).
According to OMAP3 TRM:
Set SRC(SRD) bit in MMCHS_SYSCTL register to 0x1 and wait until
it returns to 0x0.
According to OMAP4(AM335x,OMAP5,DRA7xx) TRMs, CMD(DATA) lines reset
procedure steps must be as follows:
1. Initiate CMD(DAT) line reset by writing 0x1 to SRC(SRD) bit in
MMCHS_SYSCTL register (SD_SYSCTL for AM335x).
2. Poll the SRC(SRD) bit until it is set to 0x1.
3. Wait until the SRC(SRD) bit returns to 0x0
(reset procedure is completed).
Unfortunately, at present omap_hsmmc driver has support only for
OMAP3. And as result step #2 is missing for OMAP4(AM335x,OMAP5,DRA7xx).
This sometimes leads to the fact that the waiting loop which is
required in step #3 does not executed, because SRC bit does not set
yet (at the moment of checking a condition of a loop execution).
And as a result this can cause to timeout error when sending a
next command.
In the particular case (working with eMMC witch do not respond to
some SD specific command) due to incorrect reset sequence after
command SD_CMD_SEND_IF_COND which finished with CTO flag within
64 clock cycles, the next command MMC_CMD_APP_CMD leads to a
timeout error within 1s.
So, extend CMD(DATA) lines reset procedure in func
mmc_reset_controller_fsm() by adding the missing step #2 for
OMAP4+/AM335x boards.
Kuo-Jung Su [Mon, 29 Jul 2013 05:51:43 +0000 (13:51 +0800)]
arm: dma_alloc_coherent: malloc() -> memalign()
Even though the MMU/D-cache is off, some DMA engines still
expect strict address alignment.
For example, the incoming Faraday FTMAC110 & FTGMAC100 ethernet
controllers expect the tx/rx descriptors should always be aligned
to 16-bytes boundary.
Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com> CC: Albert ARIBAUD <albert.u.boot@aribaud.net>
Tom Rini [Tue, 10 Sep 2013 13:51:44 +0000 (09:51 -0400)]
Revert "standalone-examples: support custom GCC lib"
After further testing, this patch has two problems. First,
examples/standalone/Makefile was already inherting PLATFORM_LIBS from
the top-level Makefile so this lead to duplicating the private libgcc.
Second, currently the private libgcc has a reference to 'hang' that is
not being fulfilled.
Albert ARIBAUD [Wed, 11 Sep 2013 13:52:51 +0000 (15:52 +0200)]
Merge and reformat boards.cfg and MAINTAINERS
Put all informations about targets, including state (active or
orphan) and maintainers, in boards.cfg; remove MAINTAINERS;
adjust the build system accordingly.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Signed-off-by: Philippe Reynes <tremyfr@yahoo.fr> Signed-off-by: Eric Jarrige <eric.jarrige@armadeus.org> Signed-off-by: Nicolas Colombain <nicolas.colombain@armadeus.com>
root@(none):/sys/devices/virtual/mtd/mtd0# for i in ecc_strength oobsize subpagesize; do echo $i = `cat $i`; done
ecc_strength = 16
oobsize = 224
subpagesize = 4096
The ECC strength was not properly discovered by U-Boot causing the data
written by Linux to return an -74 (EBADMSG) when read from U-Boot. This
patch fixes mxs_nand_get_ecc_strength() to function in case of a NAND
flash with page_data_size = 4096 and page_oob_size= 224.
Signed-off-by: Elie De Brauwer <eliedebrauwer@gmail.com> Acked-by: Scott Wood <scottwood@freescale.com>
arm:goni:mmc: Add sd card detection and initialization.
This change allow to use sd card on Goni the same like mmc 0.
SD card is mmc dev 1, so it can be used like this: "fatls mmc 1:2".
SD card is inited even if eMMC initialization fails.
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> CC: Minkyu Kang <mk7.kang@samsung.com> Acked-by: Jaehoon Chung <jh80.chung@samsung.com> Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
As per new convention ARCH_CPU_DTS is not defined in "dtc/Makefile".
Hence Arndale comilation is failing. Fix this by adding proper include
file in "board/samsung/dts/exynos5250-arndale.dts".
Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org> Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
York Sun [Tue, 20 Aug 2013 22:09:43 +0000 (15:09 -0700)]
powerpc/mpc85xx: Add workaround for erratum A-005125
In a very rare condition, a system hang is possible when the e500 core
initiates a guarded load to PCI / PCIe /SRIO performs a coherent write
to memory. Please refer to errata document for more details. This erratum
applies to the following SoCs and their variants, if any.
Marek Vasut [Sat, 31 Aug 2013 13:53:46 +0000 (15:53 +0200)]
ARM: mxs: Add SanDisk Sansa Fuze+ board
Add STMP3780-based Sansa Fuze+ board. This board is a small PMP
device sporting a CPU which was later rebranded to i.MX233 .
Currently supported is USB gadget mode and MMC .
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Otavio Salvador <otavio@ossystems.com.br> Cc: Stefano Babic <sbabic@denx.de>
Marek Vasut [Sat, 31 Aug 2013 13:53:45 +0000 (15:53 +0200)]
ARM: mxs: Add Creative ZEN XFi3 board
Add STMP3780-based XFi3 board. This board is a small PMP device
sporting a CPU which was later rebranded to i.MX233 . Currently
supported is USB gadget mode and both external SD and internal
Phison SD-NAND bridge .
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Otavio Salvador <otavio@ossystems.com.br> Cc: Stefano Babic <sbabic@denx.de>
Marek Vasut [Sat, 31 Aug 2013 13:53:44 +0000 (15:53 +0200)]
ARM: mxs: Receive r0 and r1 passed from BootROM
Make sure value in register r0 and r1 is preserved and passed to
the board_init_ll() and mxs_common_spl_init() where it can be
processed further. The value in r0 can be configured during the
BootStream generation to arbitary value, r1 contains pointer to
return value from CALL'd function.
This patch also clears the value in r0 before returning to BootROM
to make sure the BootROM is not confused by this value.
Finally, this patch cleans up some comments in the start.S file.
Andrew Gabbasov [Thu, 4 Jul 2013 11:27:32 +0000 (06:27 -0500)]
mx6: Fix calculation of emi_slow clock rate
This is porting of Freescale's patch from version imx_v2009.08_3.0.35_4.0.0,
that fixes the obvious mistype of bits offset macro name (ACLK_EMI_PODF_OFFSET
was used instead of ACLK_EMI_SLOW_PODF_OFFSET).
Using the occasion, change the variable name 'emi_slow_pof' to more consistent
'emi_slow_podf'.
Signed-off-by: Jason Liu <r64343@freescale.com> Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com> Acked-by: Dirk Behme <dirk.behme@de.bosch.com>
powerpc/mpc85xx: Fix the I2C bus speed error on p1022
The source clock frequency of I2C bus on p1022 is the platform(CCB)
clock, not CCB/2. The wrong source clock frequency leads to wrong
I2C bus speed setting. so, fixed it.
Stefan Roese [Mon, 26 Aug 2013 10:08:48 +0000 (12:08 +0200)]
ppc4xx: Fix GPIO handling in lwmon5 and lcd4_lwmon5 BSP
LCD4 needs a slightly different GPIO configuration than the
original LWMON5 variant. GPIO49 needs to be configured to a
default output value of 0 (permanent voltage supply).
Additionally lcd4 also needs to enable the LSB transmitter.