oweals/u-boot.git
10 years agommc: zynq: Add OF initialization support
Michal Simek [Mon, 24 Feb 2014 10:16:31 +0000 (11:16 +0100)]
mmc: zynq: Add OF initialization support

Enable initialize sdhci from DTB.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
10 years agonet: gem: Add OF initialization support
Michal Simek [Mon, 24 Feb 2014 10:16:30 +0000 (11:16 +0100)]
net: gem: Add OF initialization support

Gem can be directly initialized from DTB.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
10 years agozynq: Do not use SPL OF initialization
Michal Simek [Mon, 24 Feb 2014 10:16:29 +0000 (11:16 +0100)]
zynq: Do not use SPL OF initialization

Disable CONFIG_OF_CONTROL for SPL compilation.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
10 years agonet: emaclite: Fix OF initialization
Michal Simek [Mon, 24 Feb 2014 10:16:28 +0000 (11:16 +0100)]
net: emaclite: Fix OF initialization

- Add xilinx_emaclite_of_init to netdev.h
- Remove global data pointer from the driver
- Add better handling for error state.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
10 years agoMerge branch 'master' of git://git.denx.de/u-boot-arm
Tom Rini [Wed, 26 Feb 2014 21:49:58 +0000 (16:49 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-arm

Conflicts:
arch/arm/cpu/armv7/config.mk
board/ti/am43xx/mux.c
include/configs/am43xx_evm.h

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoarm: Switch to -mno-unaligned-access when supported by the compiler
Tom Rini [Tue, 25 Feb 2014 15:27:01 +0000 (10:27 -0500)]
arm: Switch to -mno-unaligned-access when supported by the compiler

When we tell the compiler to optimize for ARMv7 (and ARMv6 for that
matter) it assumes a default of SCTRL.A being cleared and unaligned
accesses being allowed and fast at the hardware level.  We set this bit
and must pass along -mno-unaligned-access so that the compiler will
still breakdown accesses and not trigger a data abort.

To better help understand the requirements of the project with respect
to unaligned memory access, the
Documentation/unaligned-memory-access.txt file has been added as
doc/README.unaligned-memory-access.txt and is taken from the v3.14-rc1
tag of the kernel.

Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Mans Rullgard <mans@mansr.com>
Signed-off-by: Tom Rini <trini@ti.com>
10 years agocheckpatch.pl: Add warning for new __packed additions
Tom Rini [Tue, 25 Feb 2014 15:27:00 +0000 (10:27 -0500)]
checkpatch.pl: Add warning for new __packed additions

While there are valid reasons to use __packed, often the answer is that
you should be doing something else here instead.

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoarm: remove unneeded symbol offsets and _TEXT_BASE
Albert ARIBAUD [Sat, 22 Feb 2014 16:53:43 +0000 (17:53 +0100)]
arm: remove unneeded symbol offsets and _TEXT_BASE

Remove the last uses of symbol offsets in ARM U-Boot.
Remove some needless uses of _TEXT_BASE.
Remove all _TEXT_BASE definitions.

Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
10 years agoarm: make _end compiler-generated
Albert ARIBAUD [Sat, 22 Feb 2014 16:53:42 +0000 (17:53 +0100)]
arm: make _end compiler-generated

This prevents references to _end from generating absolute
relocation records.

This change is binary invariant for ARM targets.

Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
10 years agocm_t335: fix linker file to produce full ELF
Albert ARIBAUD [Sat, 22 Feb 2014 16:53:41 +0000 (17:53 +0100)]
cm_t335: fix linker file to produce full ELF

Newly added cm_t335 was missed in commit 47ed5dd0 which
made ARM targets produce full ELF files. Fix its linker
script.

This change is binary-invariant when only .dynsym, .dynstr,
.dynamic, .plt, .interp and .gun sections are declared.

Sections .hash, .got.plt, .dynbss and .ARM.exidx are also
declared so that their (unused) content is moved out of the
u-boot binary.

Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
10 years agoRevert "ext4fs: Add ext4 extent cache for read operations"
Tom Rini [Wed, 26 Feb 2014 13:18:58 +0000 (08:18 -0500)]
Revert "ext4fs: Add ext4 extent cache for read operations"

This reverts commit fc0fc50f38a4d7d0554558076a79dfe8b0d78cd5.

The author has asked on the mailing list that we revert this for now as
it breaks write support.

Reported-by: Ɓukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Tom Rini <trini@ti.com>
10 years agoMerge branch 'master' of git://git.denx.de/u-boot-mpc85xx
Tom Rini [Tue, 25 Feb 2014 17:44:13 +0000 (12:44 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-mpc85xx

With this, fixup a trivial build error of get_effective_memsize needing
to be updated in the new board/freescale/p1010rdb/spl.c

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoarm: omap: delete unincluded omap-common/config.mk
Masahiro Yamada [Tue, 25 Feb 2014 01:01:50 +0000 (10:01 +0900)]
arm: omap: delete unincluded omap-common/config.mk

arch/arm/cpu/armv7/omap-common/config.mk is never included
because "omap-common" is not SoC name.

If we want to add OMAP-specific compiler flags,
they must be added to omap3/config.mk, omap4/config.mk, omap5/config.mk.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Tom Rini <trini@ti.com>
10 years agokbuild: Move linker sciript check to prepare1
Masahiro Yamada [Tue, 25 Feb 2014 10:26:48 +0000 (19:26 +0900)]
kbuild: Move linker sciript check to prepare1

Same as the previous commit.
Move sanity check to prepare1 target to avoid nasty troubles.

Before this commit, LDSCRIPT existence was not checked
when it was specified by CONFIG_SYS_LDSCRIPT.
Now LDSCRIPT existence is checked for all boards.

$(wildcard $(LDSCRIPT)) must point to the linker scripts
with absolute path.
Otherwise, make will terminate with a false error
on out-of-tree build.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: Fix a false error of generic board support
Masahiro Yamada [Tue, 25 Feb 2014 10:26:47 +0000 (19:26 +0900)]
kbuild: Fix a false error of generic board support

Before this commit, make terminated with an error
where is shouldn't under some condition.

This bug happened when we built a board unsupporting
generic board right after building with generic board.

For example, the following sequence failed.
(harmony uses generic board but microblaze-generic does not
support it)

  $ make harmony_config
  Configuring for harmony board...
  $ make CROSS_COMPILE=arm-linux-gnueabi-
    [ Build succeed ]
  $ make microblaze-generic_config
  Configuring for microblaze-generic board...
  $ make CROSS_COMPILE=microblaze-linux-
  Makefile:488: *** Your architecture does not support generic board.
  Please undefined CONFIG_SYS_GENERIC_BOARD in your board config file.  Stop.

We had to do "make clean" before building the microblaze board.

This commit fixes this unconvenience.

Move generic board sanity check to "prepare1" target,
which is run after generation of include/autoconf.mk.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: get back DTB at the top directory
Masahiro Yamada [Fri, 21 Feb 2014 06:34:30 +0000 (15:34 +0900)]
kbuild: get back DTB at the top directory

Commit 6ab6b2af deleted ./u-boot.dtb because it was a copy
of dts/dt.dtb.

But Simon suggested to keep u-boot.dtb at the top directory
because it is a build output.

After discussions, we agreed on revival of ./u-boot.dtb.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Tom Rini <trini@ti.com>
Acked-by: Simon Glass <sjg@chromium.org>
10 years agokbuild: add misc targets, help, ubootrelease, ubootversion
Masahiro Yamada [Mon, 24 Feb 2014 02:12:23 +0000 (11:12 +0900)]
kbuild: add misc targets, help, ubootrelease, ubootversion

"help" target has been copied from Linux 3.13 and adjusted
for U-Boot.
Unlike Linux, cleaning is done on 4 levels:
clean, clobber, mrproper, distclean.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: support building signle targets
Masahiro Yamada [Mon, 24 Feb 2014 02:12:22 +0000 (11:12 +0900)]
kbuild: support building signle targets

Import build rules of single targets from Linux 3.13.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: refactor more IMX image rules
Masahiro Yamada [Mon, 24 Feb 2014 02:12:21 +0000 (11:12 +0900)]
kbuild: refactor more IMX image rules

This commit avoids generating ./SPL twice.
  - Fist time descending to spl/
  - Second time as a prerequisite of u-boot-with-spl.imx,
    u-boot-with-nand-spl.imx.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten logs for IMX images
Masahiro Yamada [Mon, 24 Feb 2014 02:12:20 +0000 (11:12 +0900)]
kbuild: use shorten logs for IMX images

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten logs for misc targets
Masahiro Yamada [Mon, 24 Feb 2014 02:12:19 +0000 (11:12 +0900)]
kbuild: use shorten logs for misc targets

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten log for linking u-boot
Masahiro Yamada [Mon, 24 Feb 2014 02:12:18 +0000 (11:12 +0900)]
kbuild: use shorten log for linking u-boot

Move sandbox-specific link rule to arch/sandbox/config.mk.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten logs for combined rules of mkimage, padding and cat
Masahiro Yamada [Mon, 24 Feb 2014 02:12:17 +0000 (11:12 +0900)]
kbuild: use shorten logs for combined rules of mkimage, padding and cat

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten logs for cat rules
Masahiro Yamada [Mon, 24 Feb 2014 02:12:16 +0000 (11:12 +0900)]
kbuild: use shorten logs for cat rules

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten logs for mkimage rules
Masahiro Yamada [Mon, 24 Feb 2014 02:12:15 +0000 (11:12 +0900)]
kbuild: use shorten logs for mkimage rules

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: use shorten logs objcopy rules
Masahiro Yamada [Mon, 24 Feb 2014 02:12:14 +0000 (11:12 +0900)]
kbuild: use shorten logs objcopy rules

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: Add $(Q) when descending into subdirectories
Masahiro Yamada [Mon, 24 Feb 2014 02:12:13 +0000 (11:12 +0900)]
kbuild: Add $(Q) when descending into subdirectories

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Simon Glass <sjg@chromium.org>
10 years agokbuild: rename OBJCFLAGS to OBJCOPYFLAGS
Masahiro Yamada [Mon, 24 Feb 2014 02:12:12 +0000 (11:12 +0900)]
kbuild: rename OBJCFLAGS to OBJCOPYFLAGS

Rename OBJCFLAGS to OBJCOPYFLAGS beforehand to use
"cmd_objcopy" in scripts/Makefile.lib in an upcoming commit.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: move asm-offsets.h rules to ./Kbuild
Masahiro Yamada [Mon, 24 Feb 2014 02:12:11 +0000 (11:12 +0900)]
kbuild: move asm-offsets.h rules to ./Kbuild

Generate include/generated/generic-asm-offsets.h and
include/generated/asm-offsets.h in ./Kbuild.

This commit also changes the include guard.

Before this commit, __ASM_OFFSETS_H__ was used for both of them.
So we could not include generic-asm-offsets.h and asm-offsets.h
at the same time.

This commit renames the include guard of the former to
__GENERIC_ASM_OFFSETS_H__.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: abolish "depend" target
Masahiro Yamada [Mon, 24 Feb 2014 02:12:10 +0000 (11:12 +0900)]
kbuild: abolish "depend" target

This commit moves include/generated/{generic-asm-offsets.h, asm-offsets.h}
from "depend" targets to "prepare" targets and deletes "depend" targets.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agokbuild: consolidate version and timestamp headers generation
Masahiro Yamada [Mon, 24 Feb 2014 02:12:09 +0000 (11:12 +0900)]
kbuild: consolidate version and timestamp headers generation

- Generate include/generated/{timestamp.h, version.h}
   more simply by using filechk rule.

 - Add $(UBOOTRELEASE) variable and re-write u-boot.imx rule
   more simply.

 - Rename U_BOOT_VERSION in Makefile to UBOOTVERSION

   Before this commit, the same variable name, "U_BOOT_VERSION"
   was used for two different strings.

   One of them was defined in Makefile.
   It takes the form like "2014.01-rc1" and used in
   makefiles and script files.

   The other is defined in include/generated/version.h
   It takes the form like "U-Boot 2014.01-rc1-00010-gbe6d426-dirty"
   and used in C and Aseembler.

   It is confusing when grepping the source tree. So, this commit
   renames the former to UBOOTVERSION.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agoMakefile: add a new script to check -fstack-usage support
Masahiro Yamada [Wed, 11 Dec 2013 11:11:34 +0000 (20:11 +0900)]
Makefile: add a new script to check -fstack-usage support

If -fstack-usage option is given to crosstools
that do not support it, gcc displays a warning message
but still exits with status 0.

This means we can not rely on $(call cc-option,...)
to detect if -fstack-usage option is supported or not.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agofsl/usb: Limit phy_type comparison to first four characters
Nikhil Badola [Mon, 17 Feb 2014 11:28:36 +0000 (16:58 +0530)]
fsl/usb: Limit phy_type comparison to first four characters

Use first four characters for phy_type comparison. Strcmp() should not
be used to check the phy_type string which maybe parsed by hwconfig_subarg().
Hwconfig_subarg() returns part of hwconfig string starting from
phy_type value till the end of the string. Since phy_type could be
either "utmi" or "ulpi", strncmp() should be used so that a comparison
of "utmi;fsl_ddr:bank_intlv=auto" with "utmi" will succeed.

Signed-off-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agopowerpc/mpc8536DS:Increase binary size for mpc8536DS board
Haijun.Zhang [Thu, 13 Feb 2014 01:03:02 +0000 (09:03 +0800)]
powerpc/mpc8536DS:Increase binary size for mpc8536DS board

u-boot binary size for Freescale mpc8536DS platforms is 512KB.
This has been reached to upper limit of the platforms and causig
linker error. So increase the u-boot binary size to 768KB.

Signed-off-by: Haijun Zhang <Haijun.Zhang@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agoar8031/8033/phy:enable autonegotiation for ar8031/8033
Zhao Qiang [Mon, 23 Dec 2013 07:51:33 +0000 (15:51 +0800)]
ar8031/8033/phy:enable autonegotiation for ar8031/8033

Function "genphy_parse_link()" used "if (mii_reg & BMSR_ANEGCAPABLE)" before
while "if (phydev->supported & SUPPORTED_Autoneg)" now.
So assign "phydev->supported" to "phydev->drv->features" for ar8031/8033
to enable autonegotiation.

Signed-off-by: Zhao Qiang <B45475@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agopowerpc: p1010rdb: Enable p1010rdb to start from NAND/SD/SPI flash with SPL
Ying Zhang [Fri, 24 Jan 2014 07:50:09 +0000 (15:50 +0800)]
powerpc: p1010rdb: Enable p1010rdb to start from NAND/SD/SPI flash with SPL

In the previous patches, we introduced the SPL/TPL fraamework.
For SD/SPI flash booting way, we introduce the SPL to enable a loader stub. The
SPL was loaded by the code from the internal on-chip ROM. The SPL initializes
the DDR according to the SPD and loads the final uboot image into DDR, then
jump to the DDR to begin execution.

For NAND booting way, the nand SPL has size limitation on some board(e.g.
P1010RDB), it can not be more than 8KB, we can call it "minimal SPL", So the
dynamic DDR driver doesn't fit into this minimum SPL. We added the TPL that is
loaded by the the minimal SPL. The TPL initializes the DDR according to the SPD
and loads the final uboot image into DDR,then jump to the DDR to begin execution.

This patch enabled SPL/TPL for P1010RDB to support starting from NAND/SD/SPI
flash with SPL framework and initializing the DDR according to SPD in the SPL/TPL.
Because the minimal SPL load the TPL to L2 SRAM and the jump to the L2 SRAM to
execute, so the section .resetvec is no longer needed.

Signed-off-by: Ying Zhang <b40530@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agoSPL: P1022DS: fix the problem booting from spi flash
Ying Zhang [Fri, 24 Jan 2014 07:50:08 +0000 (15:50 +0800)]
SPL: P1022DS: fix the problem booting from spi flash

There was no enough memory for malloc in SPL booting from spi flash, so
relayout the memory in SPL: reduce the memory for global data from 16K
Bytes to 4K Bytes, save the space for malloc.

Signed-off-by: Ying Zhang <b40530@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agoSPL: P2020RDB: fix the problem booting from spi flash
Ying Zhang [Fri, 24 Jan 2014 07:50:07 +0000 (15:50 +0800)]
SPL: P2020RDB: fix the problem booting from spi flash

There was no enough stack in SPL, so the buffer needed in SPL is to malloc
from memory pool and to repalce the temporary variable.

Signed-off-by: Ying Zhang <b40530@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agoSPL: powerpc: expand SPL's length to 128K
Ying Zhang [Fri, 24 Jan 2014 07:50:06 +0000 (15:50 +0800)]
SPL: powerpc: expand SPL's length to 128K

1. The SPL's length of SDCARD boot has not enough,expand the SPL's
length to 128K.
2. deleted unused symbol: CONFIG_SYS_RUN_INDDR

Signed-off-by: Ying Zhang <b40530@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agopowerpc/t2081qds: Add T2081 QDS board support
Shengzhou Liu [Fri, 21 Feb 2014 05:16:19 +0000 (13:16 +0800)]
powerpc/t2081qds: Add T2081 QDS board support

T2081 QDS is a high-performance computing evaluation, development and
test platform supporting the T2081 QorIQ Power Architecture processor.

T2081QDS board Overview
-----------------------
- T2081 SoC integrating four 64-bit dual-threads e6500 cores up to 1.8GHz
- 2MB shared L2 and 512KB L3 CoreNet platform cache (CPC)
- CoreNet fabric supporting coherent and noncoherent transactions with
  prioritization and bandwidth allocation
- 32-/64-bit DDR3/DDR3LP SDRAM memory controller with ECC and interleaving
- Ethernet interfaces:
  - Two on-board 10M/100M/1G bps RGMII ports
  - Two 10Gbps XFI with on-board SFP+ cage
  - 1Gbps/2.5Gbps SGMII Riser card
  - 10Gbps XAUI Riser card
- Accelerator:
  - DPAA components consist of FMan, BMan, QMan, PME, DCE and SEC
- SerDes:
  - 8 lanes up to 10.3125GHz
  - Supports SGMII, HiGig, XFI, XAUI and Aurora debug,
- IFC:
  - 512MB NOR Flash, 2GB NAND Flash, PromJet debug port and Qixis FPGA
- eSPI:
  - Three SPI flash (16MB N25Q128A + 16MB EN25S64 + 512KB SST25WF040)
- USB:
  - Two USB2.0 ports with internal PHY (one Type-A + one micro Type mini-AB)
- PCIe:
  - Four PCI Express controllers (two PCIe 2.0 and two PCIe 3.0 with SR-IOV)
- eSDHC:
  - Supports various SD/SDHC/SDXC/eMMC devices with adapter cards and
    voltage translators
- I2C:
  - Four I2C controllers.
- UART:
  - Dual 4-pins UART serial ports

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agopowerpc/t208x: some update to support t2081
Shengzhou Liu [Tue, 21 Jan 2014 06:11:47 +0000 (14:11 +0800)]
powerpc/t208x: some update to support t2081

- fix serdes definition for t2081.
- fix clock speed for t2081.
- update ids, as CONFIG_FSL_SATA_V2 is needed only for t2080,
  T2081 has no SATA.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
10 years agoboards.cfg: Keep arc entries sorted
Fabio Estevam [Mon, 24 Feb 2014 13:32:42 +0000 (10:32 -0300)]
boards.cfg: Keep arc entries sorted

Run "tools/reformat.py -i -d '-' -s 8 <boards.cfg >boards0.cfg && mv boards0.cfg boards.cfg"
in order to keep arc entries sorted.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
10 years agoarm: delete unused macro CONFIG_ARCH_DEVICE_TREE
Masahiro Yamada [Mon, 24 Feb 2014 12:45:48 +0000 (21:45 +0900)]
arm: delete unused macro CONFIG_ARCH_DEVICE_TREE

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Tom Warren <twarren@nvidia.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Rajeshwari Birje <rajeshwari.s@samsung.com>
Cc: Inderpal Singh <inderpal.singh@linaro.org>
10 years agox86: coreboot: delete unused coreboot/config.mk
Masahiro Yamada [Mon, 24 Feb 2014 12:36:35 +0000 (21:36 +0900)]
x86: coreboot: delete unused coreboot/config.mk

HOSTCFLAGS_autoconf.mk.dep was added by commit 422322f
but it has never been used.

Cc: Vadim Bendebury <vbendeb@chromium.org>
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agopower: fix: Do not execute pmic command when not all necessary parameters are passed
Ɓukasz Majewski [Mon, 24 Feb 2014 12:33:08 +0000 (13:33 +0100)]
power: fix: Do not execute pmic command when not all necessary parameters are passed

Lack of this check resulted in a data abort when CPU tried to execute the
following command (without further mandatory input): 'pmic MAX77686_PMIC'.

Only the 'pmic list' command requires one passed parameter.
Other require at least two valid parameters for correct operation.

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
10 years agokbuild: fix SPL link bug when USE_PRIVATE_LIBGCC is "yes"
Masahiro Yamada [Mon, 24 Feb 2014 11:44:14 +0000 (20:44 +0900)]
kbuild: fix SPL link bug when USE_PRIVATE_LIBGCC is "yes"

Commit 6825a95 (kbuild: use Linux Kernel build scripts)
changed the behavior of linkage when USE_PRIAVATE_LIBGCC
is defined as "yes".
(It dropped arch/arm/lib/eabi_compat.o from the
target library.)

Affected boards are all Tegra boards.

This commit gets back the same behavior as before Kbuild series.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Tom Warren <twarren@nvidia.com>
Cc: Tom Rini <trini@ti.com>
10 years agopart_efi: fix protective mbr struct allocation
Hector Palacios [Thu, 13 Feb 2014 08:48:24 +0000 (09:48 +0100)]
part_efi: fix protective mbr struct allocation

The calloc() call was allocating space for the sizeof the struct
pointer rather than for the struct contents.
Besides, since this buffer is passed to mmc for writing and some
platforms may use cache, the legacy_mbr struct should be cache-aligned.

Signed-off-by: Hector Palacios <hector.palacios@digi.com>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
10 years agoMerge branch 'u-boot-ti/master' into 'u-boot-arm/master'
Albert ARIBAUD [Mon, 24 Feb 2014 09:29:59 +0000 (10:29 +0100)]
Merge branch 'u-boot-ti/master' into 'u-boot-arm/master'

10 years agoboard: bcm28155_ap: Add board files
Darwin Rambo [Tue, 11 Feb 2014 19:06:38 +0000 (11:06 -0800)]
board: bcm28155_ap: Add board files

Add support for the bcm28155_ap reference board.

Signed-off-by: Darwin Rambo <drambo@broadcom.com>
Reviewed-by: Steve Rae <srae@broadcom.com>
Reviewed-by: Tim Kryger <tkryger@linaro.org>
10 years agommc: kona: Add Kona mmc driver
Darwin Rambo [Tue, 11 Feb 2014 19:06:37 +0000 (11:06 -0800)]
mmc: kona: Add Kona mmc driver

Add support for the Kona SDHCI found on Broadcom mobile SoCs.

Signed-off-by: Darwin Rambo <drambo@broadcom.com>
Reviewed-by: Steve Rae <srae@broadcom.com>
Reviewed-by: Tim Kryger <tkryger@linaro.org>
10 years agoi2c: kona: Add Kona I2C driver
Darwin Rambo [Tue, 11 Feb 2014 19:06:36 +0000 (11:06 -0800)]
i2c: kona: Add Kona I2C driver

Add support for the Kona I2C controller found on Broadcom mobile SoCs.

Signed-off-by: Darwin Rambo <drambo@broadcom.com>
Reviewed-by: Steve Rae <srae@broadcom.com>
Reviewed-by: Tim Kryger <tkryger@linaro.org>
10 years agogpio: kona: Add Kona gpio driver
Darwin Rambo [Tue, 11 Feb 2014 19:06:35 +0000 (11:06 -0800)]
gpio: kona: Add Kona gpio driver

Add support for the Kona GPIO controller found on Broadcom mobile SoCs.

Signed-off-by: Darwin Rambo <drambo@broadcom.com>
Reviewed-by: Steve Rae <srae@broadcom.com>
Reviewed-by: Markus Mayer <markus.mayer@linaro.org>
Reviewed-by: Tim Kryger <tkryger@linaro.org>
10 years agoarch: bcm281xx: Initial commit of bcm281xx architecture code
Darwin Rambo [Tue, 11 Feb 2014 19:06:34 +0000 (11:06 -0800)]
arch: bcm281xx: Initial commit of bcm281xx architecture code

Add bcm281xx architecture support code including a clock framework and
chip reset.  Define register block base addresses for the bcm281xx
architecture and create an empty gpio header file required when
CONFIG_CMD_GPIO is set.

Signed-off-by: Darwin Rambo <drambo@broadcom.com>
Reviewed-by: Steve Rae <srae@broadcom.com>
Reviewed-by: Tim Kryger <tkryger@linaro.org>
10 years agoarch: kona: Initial commit of kona-common architecture code
Darwin Rambo [Tue, 11 Feb 2014 19:06:33 +0000 (11:06 -0800)]
arch: kona: Initial commit of kona-common architecture code

The Kona architecture is present on a number of Broadcom mobile SoCs
including the bcm281xx family of chips.

Signed-off-by: Darwin Rambo <drambo@broadcom.com>
Reviewed-by: Steve Rae <srae@broadcom.com>
Reviewed-by: Tim Kryger <tkryger@linaro.org>
10 years agoARM: rpi_b: rework boot scripts, enable sysboot
Stephen Warren [Fri, 14 Feb 2014 03:44:07 +0000 (20:44 -0700)]
ARM: rpi_b: rework boot scripts, enable sysboot

Rework rpi_b's bootcmd (and sub-commands) to match Tegra's bootcmd as
much as possible. This will aid in a future patch which will create a
common header e.g. config_distro_bootcmd.h.

While at it, enable booting from extlinux.conf using the sysboot command.

The iteration and componentization currently makes a little more sense
for Tegra than RPi, but I'd still like to keep the two as similar as
possible. USB and networking support is coming to the RPi in due course,
and it'll all make more sense then anyway.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
10 years agoti814x: Fix illegal use of FP ops in clock_ti814x.c
MÄns RullgÄrd [Mon, 28 Oct 2013 23:19:53 +0000 (23:19 +0000)]
ti814x: Fix illegal use of FP ops in clock_ti814x.c

The function pll_sigma_delta_val uses "float" data which is not correct.
The exact "why" of this mangling is lost to history, but this changes us
to equivalent non-FP math to get the same results.

Reported-by: Wolfgang Denk <wd@denx.de>
Acked-by: Matt Porter <mporter@linaro.org>
10 years agoomap4_common: config: remove I2C for SPL mode
Nishanth Menon [Wed, 8 Jan 2014 02:06:56 +0000 (20:06 -0600)]
omap4_common: config: remove I2C for SPL mode

Commit 6789e84ecaa8f45d053084e08c381284a04abff7 (i2c, omap24xx:
convert driver to new mutlibus/mutliadapter framework) intended to
make I2C driver compatible with latest changes. It unfortunately has
had a impact on size on SPL as well. For example on SDP4430,
32032 bytes before/MLO
35416 bytes after/MLO

With this mentioned commit, MLO stops booting on SDP4430 as only 32K
is accessible for non-secure (bootloader) s/w on GP devices and the size
increase to 56K fails boot.

On the latest u-boot commit e7be18225fbea76d1f0034b224f0d1e60f07cfcf,
MLO is now at size 35592 bytes, However, I2C is not necessary for SPL
to function as we use SR_I2C for controlling the PMIC.
Disabling I2C reduces MLO to 32224 bytes which allows
OMAP4 GP platform to boot up.

Since this is common for all OMAP4 platforms, remove the need for I2C
for SPL builds in the common config.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Sricharan R <r.sricharan@ti.com>
10 years agodra7xx_evm: Enable GPT commands
Tom Rini [Tue, 21 Jan 2014 22:15:08 +0000 (17:15 -0500)]
dra7xx_evm: Enable GPT commands

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoam43xx_evm: Enable GPT commands
Tom Rini [Tue, 21 Jan 2014 22:15:07 +0000 (17:15 -0500)]
am43xx_evm: Enable GPT commands

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoam335x_evm: Enable GPT commands
Tom Rini [Tue, 21 Jan 2014 22:15:06 +0000 (17:15 -0500)]
am335x_evm: Enable GPT commands

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoarm: omap: cm_t35: Remove CONFIG_SYS_BOOTMAPSZ to fix FDT Linux booting
Stefan Roese [Fri, 21 Feb 2014 08:28:33 +0000 (09:28 +0100)]
arm: omap: cm_t35: Remove CONFIG_SYS_BOOTMAPSZ to fix FDT Linux booting

With CONFIG_SYS_BOOTMAPSZ defined, recent Linux kernel booting
results in this error:

   Booting using the fdt blob at 0x81f00000
   Loading Kernel Image ... OK
ERROR: Failed to allocate 0x1189d bytes below 0x80004000.
device tree - allocation error
FDT creation failed! hanging...### ERROR ### Please RESET the board ###

Removing this define solves this issue.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Nikita Kiryanov <nikita@compulab.co.il>
Cc: Igor Grinberg <grinberg@compulab.co.il>
Cc: Tom Rini <trini@ti.com>
10 years agoDRA7: fix ABB efuse offset for OPP_NOM
Nishanth Menon [Tue, 18 Feb 2014 18:00:01 +0000 (12:00 -0600)]
DRA7: fix ABB efuse offset for OPP_NOM

commit 194dd74ad919e57026f385aaab7f89acf7ea79ef
(DRA7: add ABB setup for MPU voltage domain)

Made an offset typo error by using 0x4A003B24 as the efuse offset
for OPP_NOM. As per TI documentation, 0x4A003B24 is for OPP_OD, and
0x4A003B20 is for OPP_NOM. Fix the same.

Reported-by: Praveen Rao <prao@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
10 years agoARM: OMAP4: fix DDR timings for OMAP4430 ES2.0
Janne Grunau [Sun, 16 Feb 2014 20:57:18 +0000 (21:57 +0100)]
ARM: OMAP4: fix DDR timings for OMAP4430 ES2.0

DDR timings were broken since 47abc3df701d8bc26f311350aa523fc1d0f8ad4e
for PandaBoard EA1.

Signed-off-by: Janne Grunau <j@jannau.net>
10 years agoOMAP3: fix default environment for mcx board
Stefano Babic [Fri, 14 Feb 2014 11:51:27 +0000 (12:51 +0100)]
OMAP3: fix default environment for mcx board

Patch allows to load a script at the startup from MMC.

Signed-off-by: Stefano Babic <sbabic@denx.de>
10 years agoomap3: fix pinmux for mcx board
Stefano Babic [Fri, 14 Feb 2014 11:51:26 +0000 (12:51 +0100)]
omap3: fix pinmux for mcx board

New hardware version cannot recognize the SD-Card
because the SYS_NRESWARM, set as GPIO, does not
guarantee the required reset time.
Change the pin function back to enable the
SYS_NRESWARM signal driven by the SOC.

Signed-off-by: Stefano Babic <sbabic@denx.de>
10 years agoOMAP3: add missing gpio clock init and fix NAND SPL for mcx board
Stefano Babic [Fri, 14 Feb 2014 11:51:25 +0000 (12:51 +0100)]
OMAP3: add missing gpio clock init and fix NAND SPL for mcx board

Patch contains the same fixes as commit
4b9b2c300a23ca4a85811918dc92e822a9571a87
(missing CONFIG_OMAP3_GPIO_2) and commit
f9095aac793aa8917ab9b915c5d449e6dc8d3d30 (missing CONFIG_SPL_NAND_SOFTECC).
Same issues are relevant for the mcx board, too.

Signed-off-by: Stefano Babic <sbabic@denx.de>
10 years agoarm: omap3: Fix tao3530/omap3_ha SPL boot hangup (GPIO clocks not enabled)
Stefan Roese [Fri, 14 Feb 2014 08:47:17 +0000 (09:47 +0100)]
arm: omap3: Fix tao3530/omap3_ha SPL boot hangup (GPIO clocks not enabled)

Patch f33b9bd3
[arm: omap3: Enable clocks for peripherals only if they are used]
breaks SPL booting on tao3530. Since some gpio input's are
read to detect the board revision. But with this patch above, the
clocks to the GPIO subsystems are not enabled per default any more.
The GPIO banks need to be configured specifically now.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Tom Rini <trini@ti.com>
Cc: Michael Trimarchi <michael@amarulasolutions.com>
Reviewed-by: Stefano Babic <sbabic@denx.de>
10 years agoARM: AM43xx: GP-EVM: Correct GPIO used for VTT regulator control
Dave Gerlach [Mon, 10 Feb 2014 16:41:49 +0000 (11:41 -0500)]
ARM: AM43xx: GP-EVM: Correct GPIO used for VTT regulator control

Schematic indicates GPIO5_7 is to be used for VTT regulator control
rather than GPIO0_21 so modify enable_vtt_regulator to reflect this.
Without this some boards will experience DDR3 corruption and fail to
boot.

Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
[trini: Rework patch against mainline]
Signed-off-by: Tom Rini <trini@ti.com>
10 years agoAdd support for B&R KWB Motherboard
Hannes Petermaier [Fri, 7 Feb 2014 13:06:50 +0000 (14:06 +0100)]
Add support for B&R KWB Motherboard

Adds support for Bernecker & Rainer Industrieelektronik GmbH KWB
Motherboard, using TI's AM3352 SoC.

Most of code is derived from TI's AM335x_EVM

Signed-off-by: Hannes Petermaier <oe5hpm@oevsv.at>
Cc: trini@ti.com
10 years agoboard: Add support for B&R T-Series Motherboard
Hannes Petermaier [Fri, 7 Feb 2014 07:07:36 +0000 (08:07 +0100)]
board: Add support for B&R T-Series Motherboard

Adds support for Bernecker & Rainer Industrieelektronik GmbH T-Series
Motherboard, using TI's AM3352 SoC.

Most of code is derived from TI's AM335x_EVM

Signed-off-by: Hannes Petermaier <oe5hpm@oevsv.at>
10 years agoam335x: Initial support for Silica Pengwyn board
Lothar Felten [Fri, 31 Jan 2014 16:34:14 +0000 (17:34 +0100)]
am335x: Initial support for Silica Pengwyn board

This patch add support for the Silica Pengwyn board [1]
The board is based on a TI AM3354 CPU [2]
All jumpers removed it will boot from the SDcard, the console is on
UART1 accessible via the FDTI -> USB. The on board NAND flash is
supported and can act as boot medium, depending on jumper settings.
USB Host, USB Device and Ethernet are also provided but untested.

[1]
http://www.silica.com/product/silica-pengwyn-board.html
[2]
http://www.ti.com/product/am3354

Signed-off-by: Lothar Felten <lothar.felten@gmail.com>
[trini: Move CONFIG_BOARD_LATE_INIT into am335x_evm.h, drop unused
spi0_pin_mux from Pengwyn support]
Signed-off-by: Tom Rini <trini@ti.com>
10 years agoubifs: fix checkpatch warning
Karicheri, Muralidharan [Mon, 20 Jan 2014 22:10:07 +0000 (17:10 -0500)]
ubifs: fix checkpatch warning

Fix the following checkpatch warning:-

WARNING: externs should be avoided in .c files

Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
10 years agocommon, itest: pass u-boot env variables to itest.s
Heiko Schocher [Sat, 25 Jan 2014 06:27:12 +0000 (07:27 +0100)]
common, itest: pass u-boot env variables to itest.s

compare two U-Boot Environment variables with itest.s, example:

=> print tmp ver
tmp=U-Boot 2013.10-g75e
ver=U-Boot 2013.10-g75eb4bc (Jan 21 2014 - 10:35:39)MPC83XX
=> print check_ub_ver
check_ub_ver=if itest.s \${tmp} == \${ver}; then echo equal; else echo diff ;fi
=> run check_ub_ver
diff
=> setenv tmp U-Boot 2013.10-g75eb4bc (Jan 21 2014 - 10:35:39)MPC83XX
=> print tmp ver
tmp=U-Boot 2013.10-g75eb4bc (Jan 21 2014 - 10:35:39)MPC83XX
ver=U-Boot 2013.10-g75eb4bc (Jan 21 2014 - 10:35:39)MPC83XX
=> run check_ub_ver
equal

Signed-off-by: Heiko Schocher <hs@denx.de>
10 years agocommon, ubi: add ubi check volumename command
Heiko Schocher [Sat, 25 Jan 2014 06:27:11 +0000 (07:27 +0100)]
common, ubi: add ubi check volumename command

check with this ubi command, if a UBI volume with "volumename"
exists in current ubi device.

Signed-off-by: Heiko Schocher <hs@denx.de>
10 years agoboards.cfg: Delete the equivalent entries
Masahiro Yamada [Thu, 23 Jan 2014 11:09:29 +0000 (20:09 +0900)]
boards.cfg: Delete the equivalent entries

There are some entries which produce the same binaries:
 - ep8248E           is equivalent to ep8248
 - MPC8360ERDK_66    is equivalent to MPC8360ERDK
 - Adder87x/AdderUSB is equivalent to Adder
 - EVB64260_750CX    is equivalent to EVB64260

I also notice
 - Lite5200           is equivalent to icecube_5200
 - Lite5200_LOWBOOT   is equivalent to icecube_5200_LOWBOOT
 - Lite5200_LOWBOOT08 is equivalent to icecube_5200_LOWBOOT08
But I am keeping them.
(Wolfgang suggested to do so because Lite5200* are referenced
in misc documents.)

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agoboards.cfg: Place "-" in the 8th field if it is the same as 7th
Masahiro Yamada [Thu, 23 Jan 2014 11:09:28 +0000 (20:09 +0900)]
boards.cfg: Place "-" in the 8th field if it is the same as 7th

The 8th field of boards.cfg takes the form:
     <board config name>[:comma separated config options]

We should describe explicitely the 8th field only when it is necessary
to do so.
Specify "-" in the 8th field if it is the same as 7th field.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
10 years agoext4fs: Add ext4 extent cache for read operations
Ionut Nicu [Tue, 4 Feb 2014 14:48:10 +0000 (15:48 +0100)]
ext4fs: Add ext4 extent cache for read operations

In an ext4 filesystem, the inode corresponding to a file has a 60-byte
area which contains an extent header structure and up to 4 extent
structures (5 x 12 bytes).

For files that need more than 4 extents to be represented (either files
larger than 4 x 128MB = 512MB or smaller files but very fragmented),
ext4 creates extent index structures. Each extent index points to a 4KB
physical block where one extent header and additional 340 extents could
be stored.

The current u-boot ext4 code is very inefficient when it tries to load a
file which has extent indexes. For each logical file block the code will
read over and over again the same blocks of 4096 bytes from the disk.

Since the extent tree in a file is always the same, we can cache the
extent structures in memory before actually starting to read the file.

This patch creates a simple linked list of structures holding information
about all the extents used to represent a file. The list is sorted by
the logical block number (ee_block) so that we can easily find the
proper extent information for any file block.

Without this patch, a 69MB file which had just one extent index pointing
to a block with another 6 extents was read in approximately 3 minutes.
With this patch applied the same file can be read in almost 20 seconds.

Signed-off-by: Ionut Nicu <ioan.nicu.ext@nsn.com>
10 years agoSMC91111: Fix compilation warnings
Bhupesh Sharma [Wed, 5 Feb 2014 07:39:56 +0000 (13:09 +0530)]
SMC91111: Fix compilation warnings

This patch fixes the following warning messages coming out of
'drivers/net/smc91111.h' when compiled for 'vexpress_aemv8a':

warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]

Currently this issue seems to surface when SMSC is compiled for 64-bit
ARMv8 platforms, so the change is protected under CONFIG_ARM64, so that
it doesn't break other existing platforms.

In addition this patch tries to fix some checkpatch errors and warnings
(others related to camel-casing and volatile usage will be addressed
by a later patch).

This fix has been tested on both ARMv8 foundation model v1 and v2.

Signed-off-by: Bhupesh Sharma <bhupesh.sharma@freescale.com>
10 years agokbuild: fix errors of 'make *tags' and 'make cscope'
Masahiro Yamada [Fri, 21 Feb 2014 06:45:11 +0000 (15:45 +0900)]
kbuild: fix errors of 'make *tags' and 'make cscope'

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Reported-by: Nishanth Menon <nm@ti.com>
10 years agokbuild: a minor optimization of "make clobber"
Masahiro Yamada [Thu, 20 Feb 2014 08:50:57 +0000 (17:50 +0900)]
kbuild: a minor optimization of "make clobber"

The pattern rule "MLO*" can delete both MLO and MLO.byteswap.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years ago.gitignore: ignore boot images by pattern rule /u-boot*
Masahiro Yamada [Thu, 20 Feb 2014 08:43:07 +0000 (17:43 +0900)]
.gitignore: ignore boot images by pattern rule /u-boot*

U-Boot supports various boot images for various SoCs.
It is annoying to modify .gitignore file every time
we add/delete boot images.

Fortunately, there is a simple rule:
Those with file name prefix "u-boot" at the top directory
are all generated files.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
10 years agoFix memory commands for 64-bit platforms
York Sun [Wed, 12 Feb 2014 23:55:35 +0000 (15:55 -0800)]
Fix memory commands for 64-bit platforms

For aarch64, unsigned long is 64-bit data. Memory commands should be fixed
with u32 for 32-bit address access. To be clear, ushort is replace with
u16, u_char is replaced with u8.

Signed-off-by: York Sun <yorksun@freescale.com>
Acked-by: Wolfgang Denk <wd@denx.de>
10 years agopxe: prepend fdtdir to DTB name irrespective of source
Stephen Warren [Wed, 12 Feb 2014 21:30:04 +0000 (14:30 -0700)]
pxe: prepend fdtdir to DTB name irrespective of source

The directory name from an fdtdir directive in a PXE config file should
always be pre-pended to the DTB filename; it shouldn't matter whether
the DTB filename came from the $fdtfile environment variable, or whether
it was constructed dynamically from ${soc}-${board}.dtb. Fix the code to
always prepend the directory name.

Reported-by: Dennis Gilmore <dennis@ausil.us>
Fixes: c61d94d86035 ("pxe: implement fdtdir extlinux.conf tag")
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Dennis Gilmore <dennis@ausil.us>
Tested-by: Dennis Gilmore <dennis@ausil.us>
10 years agofix address of error message in mtest command
David Feng [Wed, 12 Feb 2014 08:10:08 +0000 (16:10 +0800)]
fix address of error message in mtest command

This patch deal with error message of mtest command.
When test failed, the mtest command will output error information
that include memory address and value. But the address field is
not correct or misleading.

Signed-off-by: David Feng <fenghua@phytium.com.cn>
10 years agocommon: Add get_effective_memsize() to memsize.c
York Sun [Tue, 11 Feb 2014 19:57:26 +0000 (11:57 -0800)]
common: Add get_effective_memsize() to memsize.c

This function has been around for powerpc. It is used for systems with
memory more than CONFIG_MAX_MEM_MAPPED. In case of non-contiguous memory,
this feature can limit U-boot to one block without going over the limit.

Signed-off-by: York Sun <yorksun@freescale.com>
Acked-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
10 years agodriver/ddr: Add 256 byte interleaving support
York Sun [Mon, 10 Feb 2014 21:59:44 +0000 (13:59 -0800)]
driver/ddr: Add 256 byte interleaving support

Freescale LayerScape SoCs support controller interleaving on 256 byte size.
This interleaving is mandoratory.

Signed-off-by: York Sun <yorksun@freescale.com>
10 years agoDriver/ddr: Add support of different DDR base address
York Sun [Mon, 10 Feb 2014 21:59:43 +0000 (13:59 -0800)]
Driver/ddr: Add support of different DDR base address

DDR base address has been the same from the view of core and DDR
controllers. This has changed for Freescale ARM-based SoCs. Controllers
setup DDR memory in a contiguous space and cores view it at separated
locations.

Signed-off-by: York Sun <yorksun@freescale.com>
10 years agodriver/ddr: Change Freescale ARM DDR driver to support both big and little endian
York Sun [Mon, 10 Feb 2014 21:59:42 +0000 (13:59 -0800)]
driver/ddr: Change Freescale ARM DDR driver to support both big and little endian

Initially it was believed the DDR controller on Freescale ARM would have
big endian. But some platform will have little endian.

Signed-off-by: York Sun <yorksun@freescale.com>
10 years agocommon: Remove invalid endianess conversion
Christian Eggers [Sat, 8 Feb 2014 18:27:45 +0000 (19:27 +0100)]
common: Remove invalid endianess conversion

do_bootm_standanlone() calls ntohl(images->ep) which is wrong because
endianess conversion has already been done:

do_bootm()
\-do_bootm_states()
  +-bootm_find_os()
  | \-images.ep = image_get_ep();
  |   \-uimage_to_cpu(hdr->ih_ep);
  \-boot_selected_os()
    \-do_bootm_standanlone()

Without this conversion the code works correctly at least on AT91SAM9G45.
On big endian systems there should be no difference after applying this
patch because uimage_to_cpu(x) and ntohl(x) both expand to 'x'.

Signed-off-by: Christian Eggers <ceggers@gmx.de>
10 years agoserial: arc: Convert to use default_serial_puts
Axel Lin [Sat, 8 Feb 2014 07:04:58 +0000 (15:04 +0800)]
serial: arc: Convert to use default_serial_puts

Use default_serial_puts() instead of duplicating the implementation.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
10 years agoARM: rpi_b: load /uEnv.txt from MMC at startup
Stephen Warren [Thu, 6 Feb 2014 03:49:22 +0000 (20:49 -0700)]
ARM: rpi_b: load /uEnv.txt from MMC at startup

The Pi has no flash to store an environment in the usual fashion. However,
the user may wish to customize the environment. We know that the SD card
must be present, since that's where the boot ROM has loaded U-Boot from.
So, load uEnv.txt from there early during boot. This allows the user to
e.g. customize boot_targets, in order to automatically select network
boot.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
10 years agoARM: rpi_b: convert to use distro defaults
Stephen Warren [Thu, 6 Feb 2014 03:49:21 +0000 (20:49 -0700)]
ARM: rpi_b: convert to use distro defaults

Modify the rpi_b board to include the "distro defaults" header, so that
all the config options distros expect are enabled. Remove any #defines
that enable the same options from the rpi_b.h.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
10 years agopxe: allow compilation when !defined(CONFIG_CMD_NET)
Stephen Warren [Thu, 6 Feb 2014 03:49:20 +0000 (20:49 -0700)]
pxe: allow compilation when !defined(CONFIG_CMD_NET)

pxe.c provides both the "pxe" command which relies on a network, and the
"sysboot" command which doesn't. Fix the file to compile when network
support isn't enabled. This is useful e.g. on the Raspberry Pi which has
no network support yet, but will soon support the sysboot command.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
10 years agofs/fdos: Remove
Tom Rini [Fri, 21 Feb 2014 13:42:02 +0000 (08:42 -0500)]
fs/fdos: Remove

We have an unused FAT implementation in fs/fdos, remove.

Signed-off-by: Tom Rini <trini@ti.com>
10 years agoMerge branch 'master' of git://git.denx.de/u-boot-spi
Tom Rini [Fri, 21 Feb 2014 13:00:22 +0000 (08:00 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-spi

10 years agoarc: fix relocation for big-endian target
Alexey Brodkin [Tue, 18 Feb 2014 11:10:58 +0000 (15:10 +0400)]
arc: fix relocation for big-endian target

In case of little-endian ARC700 instructions (which may include target
address) are encoded as middle-endian. That's why it's required to swap
bytes after read and ten right before write back.

But in case of big-endian ARC700 instructions are encoded as a plain
big-endian. Thus no need for byte swapping.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Francois Bedard <fbedard@synopsys.com>
Cc: Tom Rini <trini@ti.com>
cc: Noam Camus <noamc@ezchip.com>

10 years agoaxs101: change command prompt marker to match marketing name of the board
Alexey Brodkin [Sat, 8 Feb 2014 06:30:59 +0000 (10:30 +0400)]
axs101: change command prompt marker to match marketing name of the board

Proper spelling of the boad marketing name is with upper case.
So changing it from lower case to upper case here.

Cc: Tom Rini <trini@ti.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
10 years agoaxs101: fix dw gmac instantiation for updated dw gmac driver
Alexey Brodkin [Sat, 8 Feb 2014 06:24:35 +0000 (10:24 +0400)]
axs101: fix dw gmac instantiation for updated dw gmac driver

With accepted change to DW GMAC driver -
92a190aaab3536d695c51e0739f925910eb49889 (net/designware - switch driver
to phylib usage) we need to update this board because
"designware_initialize" now accepts only 2 parameters instead of 4.

Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Tom Rini <trini@ti.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
10 years agoserial/serial_arc: switch from {read|write}l to {read|write}b accessors
Alexey Brodkin [Sat, 8 Feb 2014 06:10:02 +0000 (10:10 +0400)]
serial/serial_arc: switch from {read|write}l to {read|write}b accessors

This is required for proper functionality on big-endian targets.
Memory-mapped registres of ARC UART are not 32-bit words but 8-bit bytes
so on little-endian target either acessor (_l or _b) works fine.
On big-endian only _b accessors works as expected.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Noam Camus <noamc@ezchip.com>
Cc: Tom Rini <trini@ti.com>
10 years agoserial/serial_arc: add work-around of ISS bug
Alexey Brodkin [Sat, 8 Feb 2014 06:10:01 +0000 (10:10 +0400)]
serial/serial_arc: add work-around of ISS bug

Explanation is in in-lined comment.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Noam Camus <noamc@ezchip.com>
Cc: Tom Rini <trini@ti.com>