karo-tx-uboot.git
7 years agoPrepare v2013.07-rc1 v2013.07-rc1
Tom Rini [Fri, 14 Jun 2013 15:01:39 +0000 (11:01 -0400)]
Prepare v2013.07-rc1

Signed-off-by: Tom Rini <trini@ti.com>
7 years agoMerge branch 'master' of git://git.denx.de/u-boot-spi
Tom Rini [Thu, 13 Jun 2013 19:18:35 +0000 (15:18 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-spi

7 years agoMerge branch 'master' of git://git.denx.de/u-boot-arm
Tom Rini [Thu, 13 Jun 2013 19:16:15 +0000 (15:16 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-arm

Small conflict over DRA7XX updates and adding SRAM_SCRATCH_SPACE_ADDR

Conflicts:
arch/arm/include/asm/arch-omap5/omap.h

Signed-off-by: Tom Rini <trini@ti.com>
7 years agosf: winbond: Correct the nr_blocks used for W25Q32DW
Jagannadha Sutradharudu Teki [Fri, 24 May 2013 20:43:41 +0000 (02:13 +0530)]
sf: winbond: Correct the nr_blocks used for W25Q32DW

This patch corrected the nr_blocks used for W25Q32DW SPI flash.

nr_blcoks are incorrectly assigned on below patch
"sf: winbond: add W25Q32DW"
(sha1: 772ba15474f73adc942e817cc072b6e9750836cc)

Signed-off-by: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
7 years agosf: winbond: Add support for W25Q80BW
Jagannadha Sutradharudu Teki [Mon, 27 May 2013 07:20:50 +0000 (12:50 +0530)]
sf: winbond: Add support for W25Q80BW

Add support for Winbond W25Q80BW SPI flash.

This patch corrected the flash name, nr_blocks and
also commit message header from below patch.
"sf: winbond: add W25Q32"
(sha1: c969abc47033d6f810d3c9dbdb994ea9d691d038)

Signed-off-by: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
7 years agosf: spansion: Update the name for S25FL256S flash
Jagannadha Sutradharudu Teki [Sat, 25 May 2013 17:33:11 +0000 (23:03 +0530)]
sf: spansion: Update the name for S25FL256S flash

As the per the ID tabl the flash is under Uniform 64-kB sector
architecture, hence updated with proper name.

Signed-off-by: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
7 years agospi: tegra20_sflash: Remove redundant code to set bus and cs of struct spi_slave
Axel Lin [Thu, 13 Jun 2013 08:21:42 +0000 (16:21 +0800)]
spi: tegra20_sflash: Remove redundant code to set bus and cs of struct spi_slave

It's done in spi_alloc_slave(), thus remove the redundant code.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
7 years agospi: tegra114_spi: Convert to use spi_alloc_slave()
Axel Lin [Thu, 13 Jun 2013 08:17:47 +0000 (16:17 +0800)]
spi: tegra114_spi: Convert to use spi_alloc_slave()

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
7 years agospi: armada100_spi: Remove unnecessary NULL test for dout and din
Axel Lin [Tue, 11 Jun 2013 13:57:31 +0000 (21:57 +0800)]
spi: armada100_spi: Remove unnecessary NULL test for dout and din

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Ajay Bhargav <ajay.bhargav@einfochips.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
7 years agoMerge branch 'master' of git://git.denx.de/u-boot-usb
Tom Rini [Wed, 12 Jun 2013 20:33:49 +0000 (16:33 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-usb

7 years agoarm: pxa: config option for PXA270 turbo mode
Sergey Yanovich [Tue, 21 May 2013 19:49:41 +0000 (23:49 +0400)]
arm: pxa: config option for PXA270 turbo mode

PXA270 CPU has turbo mode. The mode is 2.5 times faster than the
default run mode. Activating the mode early significantly speeds
up boot process.

Signed-off-by: Sergey Yanovich <ynvich@gmail.com>
7 years agoarm: pxa: Add support for ICP DAS LP-8x4x
Sergey Yanovich [Mon, 20 May 2013 21:26:00 +0000 (01:26 +0400)]
arm: pxa: Add support for ICP DAS LP-8x4x

LP-8x4x is a programmable automation controller by ICP DAS. It is
shipped with outdated U-Boot v1.3.0

This patch adds enough supports to boot the board:
 - 128M of 128M SDRAM
 - 32M of 48M NOR Flash memory
 - 1 of 4 Serial consoles (PXA FFUART)
 - 2 of 2 Ethernet controllers (DM9000)

Signed-off-by: Sergey Yanovich <ynvich@gmail.com>
Series-to: u-boot
Series-cc: marex

7 years agousb, composite: after unregister gadget driver set composite to NULL
Heiko Schocher [Tue, 4 Jun 2013 09:21:32 +0000 (11:21 +0200)]
usb, composite: after unregister gadget driver set composite to NULL

Without this, second usb_composite_register() call fails always
with -EINVAL.

Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Marek Vasut <marex@denx.de>
7 years agousb: ehci: add missing cache managment
Stephen Warren [Fri, 24 May 2013 21:03:17 +0000 (15:03 -0600)]
usb: ehci: add missing cache managment

Commit 8f62ca6 "usb: ehci: Support interrupt transfers via periodic list"
didn't include any cache management in the new interrupt transfer path.
It also added an extra write to or_asynclistaddr in usb_lowlevel_init(),
without having flushed out the data there.

Add the missing cache management calls, so that the code works again.

This allows the USB keyboard on Tegra's Seaboard/Springbank boards to
work.

Cc: Patrick Georgi <patrick@georgi-clan.de>
Cc: Vincent Palatin <vpalatin@chromium.org>
Cc: Julius Werner <jwerner@chromium.org>
Cc: Simon Glass <sjg@chromium.org>
Cc: Marek Vasut <marex@denx.de>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
7 years agousb: gadget: add Faraday FOTG210 USB gadget support
Kuo-Jung Su [Wed, 15 May 2013 07:29:24 +0000 (15:29 +0800)]
usb: gadget: add Faraday FOTG210 USB gadget support

The Faraday FOTG210 is an OTG chip which could operate
as either an EHCI Host or a USB Device at a time.

Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com>
CC: Marek Vasut <marex@denx.de>
7 years agousb: ehci: add Faraday USB 2.0 EHCI support
Kuo-Jung Su [Wed, 15 May 2013 07:29:23 +0000 (15:29 +0800)]
usb: ehci: add Faraday USB 2.0 EHCI support

This patch adds support to both Faraday FUSBH200 and FOTG210,
the differences between Faraday EHCI and standard EHCI are
listed bellow:

1. The PORTSC starts at 0x30 instead of 0x44.
2. The CONFIGFLAG(0x40) is not only un-implemented, and
   also has its address space removed.
3. Faraday EHCI is a TDI design, but it doesn't
   compatible with the general TDI implementation
   found at both U-Boot and Linux.
4. The ISOC descriptors differ from standard EHCI in
   several ways. But since U-boot doesn't support ISOC,
   we don't have to worry about that.

Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com>
CC: Marek Vasut <marex@denx.de>
7 years agousb: hub: make minimum power-on delay configurable
Kuo-Jung Su [Wed, 15 May 2013 07:29:22 +0000 (15:29 +0800)]
usb: hub: make minimum power-on delay configurable

This patch makes the minimum power-on delay for USB HUB
become configurable. The original design waits at least
100 msec here, but some EHCI controlers(e.g. Faraday EHCI)
are known to require much longer delay interval.

Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com>
CC: Marek Vasut <marex@denx.de>
7 years agousb: ehci: add weak-aliased function for PORTSC
Kuo-Jung Su [Wed, 15 May 2013 07:29:21 +0000 (15:29 +0800)]
usb: ehci: add weak-aliased function for PORTSC

There is at least one non-EHCI compliant controller (i.e. Faraday EHCI)
not only leave RESERVED and CONFIGFLAG registers un-implemented
but also has their address spaces removed.

As an result, the PORTSC register of Faraday EHCI always
starts from 0x30 instead of 0x44 in standard EHCI.

So that we'll need a weak-aliased function for abstraction.

Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com>
CC: Marek Vasut <marex@denx.de>
7 years agousb: ehci: prevent bad PORTSC register access
Kuo-Jung Su [Wed, 15 May 2013 07:29:20 +0000 (15:29 +0800)]
usb: ehci: prevent bad PORTSC register access

1. The 'index' of ehci_submit_root() is not always > 0.

   e.g.
   While it gets invoked from usb_get_descriptor(),
   the 'index' is always a '0'. (See ch.9 of USB2.0)

2. The PORTSC register is not always required, and thus it
   should only report a port error when necessary.
   It would cause a port scan failure if the ehci_submit_root()
   always gets terminated by a port error.

Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com>
CC: Marek Vasut <marex@denx.de>
7 years agousb: gadget: Use unaligned access for wMaxPacketSize
Vivek Gautam [Mon, 13 May 2013 10:23:38 +0000 (15:53 +0530)]
usb: gadget: Use unaligned access for wMaxPacketSize

Use get_unaligned() while fetching wMaxPacketSize to avoid
voilating any alignment rules.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Piotr Wilczek <p.wilczek@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Lukasz Dalek <luk0104@gmail.com>
Cc: Marek Vasut <marex@denx.de>
7 years agousb: Use get_unaligned() in usb_endpoint_maxp() for wMaxPacketSize
Vivek Gautam [Mon, 13 May 2013 10:23:37 +0000 (15:53 +0530)]
usb: Use get_unaligned() in usb_endpoint_maxp() for wMaxPacketSize

Use unaligned access to fetch wMaxPacketSize in usb_endpoint_maxp()
api.
In its absence we see following data abort message:
==============================================================
data abort

    MAYBE you should read doc/README.arm-unaligned-accesses

pc : [<bf794e24>]          lr : [<bf794e1c>]
sp : bf37c7b0  ip : 0000002f     fp : 00000000
r10: 00000000  r9 : 00000002     r8 : bf37fecc
r7 : 00000001  r6 : bf7d8931     r5 : bf7d891c  r4 : bf7d8800
r3 : bf7d65b0  r2 : 00000002     r1 : bf7d65b4  r0 : 00000027
Flags: nZCv  IRQs off  FIQs off  Mode SVC_32
Resetting CPU ...

resetting ...
==============================================================

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Cc: Ilya Yanok <ilya.yanok@cogentembedded.com>
Cc: Marek Vasut <marex@denx.de>
7 years agousb: asix: Move software resets to basic_init
Julius Werner [Sat, 11 May 2013 20:35:02 +0000 (13:35 -0700)]
usb: asix: Move software resets to basic_init

The ASIX driver calls a basic_init() function during get_info(), so that
not all initialization tasks need to be redone on every init().
Unfortunately, the most important one is still triggered too often: the
driver does a full port and MII reset on every asix_init(), requiring up
to several seconds to reestablish the link.

This patch confines that software reset into the asix_basic_init()
function so that it will only be executed once. This saves about a
second of boot time on systems using BOOTP.

Note: this patch was previously submitted many moons ago as:

   usb: usbeth: asix: Do a fast init if link already established

That patch seens to have been lost or forgotten, so this is a rebased
version. It is tested on snow with a Asix USB dongle (Cisco).

Signed-off-by: Julius Werner <jwerner@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
7 years agousb: Correct CLEAR_FEATURE code in ehci-hcd
Simon Glass [Sat, 11 May 2013 02:49:00 +0000 (19:49 -0700)]
usb: Correct CLEAR_FEATURE code in ehci-hcd

This commit broke USB2 on link (Chromebook Pixel):

  020bbcb usb: hub: Power-cycle on root-hub ports

However the root cause seems to be a missing mask and missing 'break'
in ehci-hcd.c. This patch fixes both.

On link, 'usb start' with a USB keyboard and memory stick inserted now
finds both. The keyboard works as expected. Also ext2ls shows a directory
listing from the memory stick.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agousb: workaround non-working keyboards.
Vincent Palatin [Sat, 11 May 2013 02:48:59 +0000 (19:48 -0700)]
usb: workaround non-working keyboards.

If the USB keyboard is not answering properly the first request on its
interrupt endpoint, just skip it and try the next one.

This workarounds an issue with a wireless mouse dongle which presents
itself both as a keyboard and a mouse but has a non-functional keyboard
interface.

Signed-off-by: Vincent Palatin <vpalatin@chromium.org>
(cherry picked from commit 012bbf0ce0301be2482857e3f03b481dd15c2340)
Rebased to upstream/master:
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Vincent Palatin <vpalatin@chromium.org>
7 years agousb: properly re-initialize the USB keyboard.
Vincent Palatin [Sat, 11 May 2013 02:48:58 +0000 (19:48 -0700)]
usb: properly re-initialize the USB keyboard.

Allow to reconfigure properly the USB keyboard driver when we enumerate
several times the USB devices and its position in the device tree has
changes.

Signed-off-by: Vincent Palatin <vpalatin@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Vincent Palatin <vpalatin@chromium.org>
7 years agoMerge branch 'master' of git://git.denx.de/u-boot-74xx-7xx
Tom Rini [Tue, 11 Jun 2013 22:11:47 +0000 (18:11 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-74xx-7xx

7 years agoppc: ppmc7xx: Fix possible out-of-bound access
Marek Vasut [Mon, 20 May 2013 03:01:40 +0000 (05:01 +0200)]
ppc: ppmc7xx: Fix possible out-of-bound access

The flash_info_t->start[] field is limited in size by CONFIG_SYS_MAX_FLASH_SECT
macro, which is set to 19 for this board in the board config file. If we inspect
the board/ppmc7xx/flash.c closely, especially the flash_get_size() function, we
can notice the "switch ((long)flashtest)" at around line 80 having a few results
which will set flash_info_t->sector_count to value higher than 19, for example
"case AMD_ID_LV640U" will set it to 128. Notice that right underneath, iteration
over flash_info_t->start[] happens and the upper bound for the interation is
flash_info_t->sector_count. Now if the sector_count is 128 as it is for the
AMD_ID_LV640U case, but the CONFIG_SYS_MAX_FLASH_SECT limiting the start[] is
only 19, an access past the start[] array much happen. Moreover, during this
iteration, the field is written to, so memory corruption is inevitable.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Tom Rini <trini@ti.com>
Cc: Richard Danter <richard.danter@windriver.com>
7 years agopowerpc: fix 8xx and 82xx type-punning warnings with GCC 4.7
Scott Wood [Sat, 18 May 2013 01:01:54 +0000 (20:01 -0500)]
powerpc: fix 8xx and 82xx type-punning warnings with GCC 4.7

C99's strict aliasing rules are insane to use in low-level code such as a
bootloader, but as Wolfgang has rejected -fno-strict-aliasing in the
past, add a union so that 16-bit accesses can be performed.

Compile-tested only.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Acked-by: Wolfgang Denk <wd@denx.de>
7 years agocosmetic: arm: fix comments in arch/arm/lib/crt0.S
Masahiro Yamada [Wed, 15 May 2013 08:33:16 +0000 (17:33 +0900)]
cosmetic: arm: fix comments in arch/arm/lib/crt0.S

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
7 years agoMerge branch 'u-boot-ti/master' into 'u-boot-arm/master'
Albert ARIBAUD [Mon, 10 Jun 2013 16:28:37 +0000 (18:28 +0200)]
Merge branch 'u-boot-ti/master' into 'u-boot-arm/master'

7 years agoarm/km: make local functions static
Holger Brunck [Mon, 6 May 2013 13:04:51 +0000 (15:04 +0200)]
arm/km: make local functions static

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
7 years agoarm: da830: moved pinmux configurations to the arch tree
Vishwanathrao Badarkhe, Manish [Wed, 29 May 2013 21:55:11 +0000 (21:55 +0000)]
arm: da830: moved pinmux configurations to the arch tree

Move pinmux configurations for the DA830 SoCs from board file
to the arch tree so that it can be used for all da830 based devices.
Also, avoids duplicate pinmuxing in case of NAND.

Signed-off-by: Vishwanathrao Badarkhe, Manish <manishv.b@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>
Acked-by: Christian Riesch <christian.riesch@omicron.at>
7 years agoARM: DRA7: Add Maintainer
Lokesh Vutla [Fri, 7 Jun 2013 00:59:02 +0000 (00:59 +0000)]
ARM: DRA7: Add Maintainer

Adding Maintainer for DRA7xx.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoOMAP5: Enable access to auxclk registers
Lubomir Popov [Wed, 15 May 2013 04:41:01 +0000 (04:41 +0000)]
OMAP5: Enable access to auxclk registers

auxclk0 and auxclk1 are utilized on some OMAP5 boards.
Define the infrastructure needed for accessing them
without using magic numbers.

Also remove unrelated TPS62361 defines from clocks.h

Signed-off-by: Lubomir Popov <lpopov@mm-sol.com>
7 years agoARM: OMAP: I2C: New read, write and probe functions
Lubomir Popov [Sat, 1 Jun 2013 06:44:38 +0000 (06:44 +0000)]
ARM: OMAP: I2C: New read, write and probe functions

New i2c_read, i2c_write and i2c_probe functions, tested on OMAP4
(4430/60/70), OMAP5 (5430) and AM335X (3359); should work on older
OMAPs and derivatives as well. The only anticipated exception would
be the OMAP2420, which shall require driver modification.
- Rewritten i2c_read to operate correctly with all types of chips
  (old function could not read consistent data from some I2C slaves).
- Optimised i2c_write.
- New i2c_probe, performs write access vs read. The old probe could
  hang the system under certain conditions (e.g. unconfigured pads).
- The read/write/probe functions try to identify unconfigured bus.
- Status functions now read irqstatus_raw as per TRM guidelines
  (except for OMAP243X and OMAP34XX).
- Driver now supports up to I2C5 (OMAP5).

Signed-off-by: Lubomir Popov <lpopov@mm-sol.com>
Tested-by: Heiko Schocher <hs@denx.de>
7 years agoarm: Remove OMAP2420H4 and all omap24xx support
Tom Rini [Tue, 4 Jun 2013 12:02:06 +0000 (12:02 +0000)]
arm: Remove OMAP2420H4 and all omap24xx support

The omap2420H4 was the only mainline omap24xx board.  Prior to being
fixed by Jon Hunter in time for v2013.04 it had been functionally broken
for a very long time.  Remove this board as there's not been interest in
it in U-Boot for quite a long time.

Signed-off-by: Tom Rini <trini@ti.com>
7 years agoda830: add MMC support
Vishwanathrao Badarkhe, Manish [Wed, 22 May 2013 03:38:48 +0000 (03:38 +0000)]
da830: add MMC support

Add MMC support for da830 boards in order to perform
mmc operations(read,write and erase).

Signed-off-by: Vishwanathrao Badarkhe, Manish <manishv.b@ti.com>
7 years agoARM: OMAP5: Power: Add more functionality to Palmas driver
Lubomir Popov [Thu, 6 Jun 2013 04:16:40 +0000 (04:16 +0000)]
ARM: OMAP5: Power: Add more functionality to Palmas driver

Add some useful functions, and the corresponding definitions.

Add support for powering on the dra7xx_evm SD/MMC LDO
(courtesy Lokesh Vutla <lokeshvutla@ti.com>).

Signed-off-by: Lubomir Popov <lpopov@mm-sol.com>
Reviewed-by: Tom Rini <trini@ti.com>
7 years agoARM: DRA7xx: EMIF: Change settings required for EVM board
Sricharan R [Thu, 30 May 2013 03:19:39 +0000 (03:19 +0000)]
ARM: DRA7xx: EMIF: Change settings required for EVM board

DRA7 EVM board has the below configuration. Adding the
settings for the same here.

   2Gb_1_35V_DDR3L part * 2 on EMIF1
   2Gb_1_35V_DDR3L part * 4 on EMIF2

Signed-off-by: Sricharan R <r.sricharan@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: clocks: Update PLL values
Lokesh Vutla [Thu, 30 May 2013 03:19:38 +0000 (03:19 +0000)]
ARM: DRA7xx: clocks: Update PLL values

Update PLL values.
SYS_CLKSEL value for 20MHz is changed to 2. In other platforms
SYS_CLKSEL value 2 represents reserved. But in sys_clk array
ind 1 is used for 13Mhz. Since other platforms are not using
13Mhz, reusing index 1 for 20MHz.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Sricharan R <r.sricharan@ti.com>
7 years agoARM: DRA7xx: Update pinmux data
Lokesh Vutla [Thu, 30 May 2013 03:19:37 +0000 (03:19 +0000)]
ARM: DRA7xx: Update pinmux data

Updating pinmux data as specified in the latest DM

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Balaji T K <balajitk@ti.com>
7 years agommc: omap_hsmmc: Update pbias programming
Balaji T K [Thu, 6 Jun 2013 05:04:32 +0000 (05:04 +0000)]
mmc: omap_hsmmc: Update pbias programming

Update pbias programming sequence for OMAP5 ES2.0/DRA7

Signed-off-by: Balaji T K <balajitk@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: Correct SRAM END address
Sricharan R [Thu, 30 May 2013 03:19:35 +0000 (03:19 +0000)]
ARM: DRA7xx: Correct SRAM END address

NON SECURE SRAM is 512KB in DRA7xx devices.
So fixing it here.

Signed-off-by: Sricharan R <r.sricharan@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: Correct the SYS_CLK to 20MHZ
Sricharan R [Thu, 30 May 2013 03:19:34 +0000 (03:19 +0000)]
ARM: DRA7xx: Correct the SYS_CLK to 20MHZ

The sys_clk on the dra evm board is 20MHZ.
Changing the configuration for the same.
And also moving V_SCLK, V_OSCK defines to
arch/clock.h for OMAP4+ boards.

Signed-off-by: Sricharan R <r.sricharan@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: Change the Debug UART to UART1
Sricharan R [Thu, 30 May 2013 03:19:33 +0000 (03:19 +0000)]
ARM: DRA7xx: Change the Debug UART to UART1

Serial UART is connected to UART1. So add the change
for the same.

Signed-off-by: Sricharan R <r.sricharan@ti.com>
7 years agoARM: DRA7xx: Do not enable srcomp for DRA7xx Soc's
Lokesh Vutla [Thu, 30 May 2013 03:19:32 +0000 (03:19 +0000)]
ARM: DRA7xx: Do not enable srcomp for DRA7xx Soc's

Slew rate compensation cells are not present for DRA7xx
Soc's. So return from function srcomp_enable() if soc is not
OMAP54xx.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: OMAP5: DRA7xx: support class 0 optimized voltages
Nishanth Menon [Thu, 30 May 2013 03:19:31 +0000 (03:19 +0000)]
ARM: OMAP5: DRA7xx: support class 0 optimized voltages

DRA752 now uses AVS Class 0 voltages which are voltages in efuse.

This means that we can now use the optimized voltages which are
stored as mV values in efuse and program PMIC accordingly.

This allows us to go with higher OPP as needed in the system without
the need for implementing complex AVS logic.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: clocks: Fixing i2c_init for PMIC
Lokesh Vutla [Thu, 30 May 2013 03:19:30 +0000 (03:19 +0000)]
ARM: DRA7xx: clocks: Fixing i2c_init for PMIC

In DRA7xx Soc's voltage scaling is done using GPI2C.
So i2c_init should happen before scaling. I2C driver
uses __udelay which needs timer to be initialized.
So moving timer_init just before voltage scaling.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: power Add support for tps659038 PMIC
Lokesh Vutla [Thu, 30 May 2013 03:19:29 +0000 (03:19 +0000)]
ARM: DRA7xx: power Add support for tps659038 PMIC

TPS659038 is the power IC used in DRA7XX boards.
Adding support for this and also adding pmic data
for DRA7XX boards.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: DRA7xx: Add control id code for DRA7xx
Lokesh Vutla [Thu, 30 May 2013 03:19:28 +0000 (03:19 +0000)]
ARM: DRA7xx: Add control id code for DRA7xx

The registers that are used for device identification
are changed from OMAP5 to DRA7xx.
Using the correct registers for DRA7xx.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: OMAP4+: pmic: Make generic bus init and write functions
Lokesh Vutla [Thu, 30 May 2013 02:54:33 +0000 (02:54 +0000)]
ARM: OMAP4+: pmic: Make generic bus init and write functions

Voltage scaling can be done in two ways:
-> Using SR I2C
-> Using GP I2C
In order to support both, have a function pointer in pmic_data
so that we can call as per our requirement.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: OMAP2+: Rename asm/arch/clocks.h asm/arch/clock.h
Lokesh Vutla [Thu, 30 May 2013 02:54:32 +0000 (02:54 +0000)]
ARM: OMAP2+: Rename asm/arch/clocks.h asm/arch/clock.h

To be consistent with other ARM platforms,
renaming asm/arch-omap*/clocks.h to asm/arch-omap*/clock.h

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: OMAP5: clocks: Do not enable sgx clocks
Sricharan R [Thu, 30 May 2013 02:54:31 +0000 (02:54 +0000)]
ARM: OMAP5: clocks: Do not enable sgx clocks

SGX clocks should be enabled only for OMAP5 ES1.0.
So this can be removed.

Signed-off-by: Sricharan R <r.sricharan@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoARM: OMAP4+: Cleanup header files
Lokesh Vutla [Thu, 30 May 2013 02:54:30 +0000 (02:54 +0000)]
ARM: OMAP4+: Cleanup header files

After having the u-boot clean up series, there are
many definitions that are unused in header files.
Removing all those unused ones.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoOMAP5: Fix bug in omap5_es1_prcm struct
Lubomir Popov [Sun, 26 May 2013 10:03:17 +0000 (10:03 +0000)]
OMAP5: Fix bug in omap5_es1_prcm struct

The newly introduced function setup_warmreset_time(), called
from within prcm_init(), tries to write to the prm_rsttime
OMAP5 register. The struct member holding this register's
address is however initialized for OMAP5 ES2.0 only. On ES1.0
devices this uninitialized value causes a second (warm) reset
at startup.

Add .prm_rsttime address init to the ES1.0 struct.

Signed-off-by: Lubomir Popov <lpopov@mm-sol.com>
Acked-by: Tom Rini <trini@ti.com>
7 years agoOMAP5: add ABB setup for MPU voltage domain
Andrii Tseglytskyi [Mon, 20 May 2013 22:42:09 +0000 (22:42 +0000)]
OMAP5: add ABB setup for MPU voltage domain

Patch adds a call of abb_setup() function, and proper registers
definitions needed for ABB setup sequence. ABB is initialized
for MPU voltage domain.

Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@ti.com>
7 years agoOMAP3+: introduce generic ABB support
Andrii Tseglytskyi [Mon, 20 May 2013 22:42:08 +0000 (22:42 +0000)]
OMAP3+: introduce generic ABB support

Adaptive Body Biasing (ABB) modulates transistor bias voltages
dynamically in order to optimize switching speed versus leakage.
Adaptive Body-Bias ldos are present for some voltage domains
starting with OMAP3630. There are three modes of operation:

* Bypass - the default, it just follows the vdd voltage
* Foward Body-Bias - applies voltage bias to increase transistor
  performance at the cost of power.  Used to operate safely at high
  OPPs.
* Reverse Body-Bias - applies voltage bias to decrease leakage and
  save power.  Used to save power at lower OPPs.

Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@ti.com>
Acked-by: Nishanth Menon <nm@ti.com>
7 years agoam33xx: Board: Make CPSW section of ethernet initialization depend on CPSW driver
Joel A Fernandes [Tue, 7 May 2013 05:52:55 +0000 (05:52 +0000)]
am33xx: Board: Make CPSW section of ethernet initialization depend on CPSW driver

Not doing so breaks cases where CPSW is not required such as for USB RNDIS network boot.

Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
7 years agoMIPS: asm/errno.h: switch to asm-generic/errno.h
Daniel Schwierzeck [Sat, 8 Jun 2013 17:23:12 +0000 (19:23 +0200)]
MIPS: asm/errno.h: switch to asm-generic/errno.h

This fixes several warnings like

In file included from ./u-boot/include/linux/mtd/mtd.h:13:0,
                 from env_onenand.c:37:
./u-boot/build/vct_platinumavc_onenand_small/include2/asm/errno.h:52:0: warning: "ENOMSG" redefined [enabled by default]

Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
7 years agoMIPS: fix __raw_* IO accessors
Gabor Juhos [Wed, 20 Feb 2013 22:03:01 +0000 (22:03 +0000)]
MIPS: fix __raw_* IO accessors

The purpose of the __raw* IO accessors is to provide
IO access in native-endian order. However in the current
MIPS implementation, the 16 and 32 bit variants of the
__raw accessors are swapping the values on big-endian
systems if the CONFIG_SWAP_IO_SPACE option is enabled.

The patch changes the IO accessor macros to fix this
broken behaviour.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Cc: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
7 years agoMerge branch 'u-boot-imx/master' into 'u-boot-arm/master'
Albert ARIBAUD [Sat, 8 Jun 2013 12:35:10 +0000 (14:35 +0200)]
Merge branch 'u-boot-imx/master' into 'u-boot-arm/master'

Conflicts:
drivers/serial/Makefile

7 years agopci: introduce CONFIG_PCI_INDIRECT_BRIDGE option
Gabor Juhos [Thu, 30 May 2013 07:06:12 +0000 (07:06 +0000)]
pci: introduce CONFIG_PCI_INDIRECT_BRIDGE option

The pci_indirect.c file is always compiled when
CONFIG_PCI is defined although the indirect PCI
bridge support is not needed by every board.

Introduce a new CONFIG_PCI_INDIRECT_BRIDGE
config option and only compile indirect PCI
bridge support if this options is enabled.

Also add the new option into the configuration
files of the boards which needs that.

Compile tested for powerpc, x86, arm and nds32.
MAKEALL results:

powerpc:
  --------------------- SUMMARY ----------------------------
  Boards compiled: 641
  Boards with warnings but no errors: 2 ( ELPPC MPC8323ERDB )
  ----------------------------------------------------------
  Note: the warnings for ELPPC and MPC8323ERDB are present even
  without the actual patch.

x86:
  --------------------- SUMMARY ----------------------------
  Boards compiled: 1
  ----------------------------------------------------------

arm:
  --------------------- SUMMARY ----------------------------
  Boards compiled: 311
  ----------------------------------------------------------

nds32:
  --------------------- SUMMARY ----------------------------
  Boards compiled: 3
  ----------------------------------------------------------

Cc: Tom Rini <trini@ti.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
7 years agofdt: remove unaligned access in fdt_fixup_ethernet()
Stephen Warren [Mon, 27 May 2013 18:01:19 +0000 (18:01 +0000)]
fdt: remove unaligned access in fdt_fixup_ethernet()

Some ARM compilers may emit code that makes unaligned accesses when
faced with constructs such as:

char mac[16] = "ethaddr";

Replace this with a strcpy() call instead to avoid this. strcpy() is
used here, rather than replacing all usage of the mac variable with the
string itself, since the loop itself sprintf()s to the variable each
iteration, so strcpy() is doing basically the same thing.

Reported-by: Florian Meier
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
7 years agogeneric_board: reduce the redundancy of gd_t struct members
Masahiro Yamada [Mon, 27 May 2013 00:37:30 +0000 (00:37 +0000)]
generic_board: reduce the redundancy of gd_t struct members

This commit refactors common/board_f.c and common/board_r.c
in order to delete the dest_addr and dest_addr_sp from
gd_t struct.

As mentioned as follows in include/asm-generic/global_data.h,

  /* TODO: is this the same as relocaddr, or something else? */
  unsigned long dest_addr;        /* Post-relocation address of U-Boot */

dest_addr is the same as relocaddr.
Likewise, dest_addr_sp is the same as start_addr_sp.

It seemed dest_addr/dest_addr_sp was used only as a scratch variable
to calculate relocaddr/start_addr_sp, respectively.

With a little refactoring, we can delete dest_addr and dest_addr_sp.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Simon Glass <sjg@chromium.org>
7 years agoam335x: enable falcon boot mode for mmc (raw and fat) and nand
Peter Korsgaard [Mon, 13 May 2013 08:36:30 +0000 (08:36 +0000)]
am335x: enable falcon boot mode for mmc (raw and fat) and nand

Jump into full u-boot mode if a 'c' character is received on the uart.

We need to adjust the spl bss/malloc area to not overlap with the
loadaddr of the kernel (sdram + 32k), so move it past u-boot instead.

For raw mmc, we store the kernel parameter area in the free space after
the MBR (if used). For nand, we use the last sector of the partition
reserved for u-boot.

This also enables the spl command in the full u-boot so the kernel
parameter area snapshot can be created.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
7 years agospl_mmc: add Falcon mode support for raw variant
Peter Korsgaard [Mon, 13 May 2013 08:36:29 +0000 (08:36 +0000)]
spl_mmc: add Falcon mode support for raw variant

If Falcon mode support is enabled (and the system isn't directed into
booting u-boot), it will instead try to load kernel from sector
CONFIG_SYS_MMCSD_RAW_MODE_KERNEL_SECTOR and
CONFIG_SYS_MMCSD_RAW_MODE_ARGS_SECTORS of kernel argument parameters
starting from sector CONFIG_SYS_MMCSD_RAW_MODE_ARGS_SECTOR.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
7 years agospl_mmc: mmc_load_image_raw(): Add sector argument
Peter Korsgaard [Mon, 13 May 2013 08:36:28 +0000 (08:36 +0000)]
spl_mmc: mmc_load_image_raw(): Add sector argument

So we can use it for falcon mode as well.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
7 years agospl_mmc: add Falcon mode support for FAT variant
Peter Korsgaard [Mon, 13 May 2013 08:36:27 +0000 (08:36 +0000)]
spl_mmc: add Falcon mode support for FAT variant

If Falcon mode support is enabled (and the system isn't directed into
booting u-boot), it will instead try to load kernel from
CONFIG_SPL_FAT_LOAD_KERNEL_NAME file and kernel argument parameters from
CONFIG_SPL_FAT_LOAD_ARGS_NAME, both from the same partition as u-boot.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
7 years agodevkit8000: Add SPL_OS for MMC support
Tom Rini [Fri, 7 Jun 2013 18:16:43 +0000 (14:16 -0400)]
devkit8000: Add SPL_OS for MMC support

Signed-off-by: Tom Rini <trini@ti.com>
7 years agospl_mmc: mmc_load_image_fat(): Add filename argument and move fat init out
Peter Korsgaard [Mon, 13 May 2013 08:36:26 +0000 (08:36 +0000)]
spl_mmc: mmc_load_image_fat(): Add filename argument and move fat init out

So we can use it for falcon mode as well.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
7 years agospl_mmc: return error from mmc_load_image_{raw, fat} rather than hanging
Peter Korsgaard [Mon, 13 May 2013 08:36:25 +0000 (08:36 +0000)]
spl_mmc: return error from mmc_load_image_{raw, fat} rather than hanging

So we can instead fallback to doing something else on errors.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
7 years agoMerge branch 'master' of git://git.denx.de/u-boot-video
Tom Rini [Fri, 7 Jun 2013 12:35:36 +0000 (08:35 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-video

7 years agoMerge branch 'master' of git://git.denx.de/u-boot-nand-flash
Tom Rini [Fri, 7 Jun 2013 12:34:34 +0000 (08:34 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-nand-flash

7 years agoARM: tegra: only enable SCU on Tegra20
Tom Warren [Thu, 23 May 2013 12:26:18 +0000 (12:26 +0000)]
ARM: tegra: only enable SCU on Tegra20

The non-SPL build of U-Boot on Tegra only runs on a single CPU, and
hence there is no need to enable the SCU when running U-Boot. If an
SMP OS is booted, and it needs the SCU enabled, it will enable the SCU
itself. U-Boot doing so is redundant.

The one exception is Tegra20, where an enabled SCU is required for some
aspects of PCIe to work correctly.

Some Tegra SoCs contain CPUs without a software-controlled SCU. In this
case, attempting to turn it on actively causes problems. This is the case
for Tegra114. For example, when running Linux, the first (or at least
some very early) user-space process will trigger the following kernel
message:

Unhandled fault: imprecise external abort (0x406) at 0x00000000

This is typically accompanied by that process receving a fatal signal,
and exiting. Since this process is usually pid 1, this causes total
system boot failure.

Signed-off-by: Tom Warren <twarren@nvidia.com>
[swarren, fleshed out description, ported to upstream chipid APIs]
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
7 years agovf610twr: Drop unneeded 'status' variable
Fabio Estevam [Wed, 5 Jun 2013 11:34:48 +0000 (11:34 +0000)]
vf610twr: Drop unneeded 'status' variable

No need to use the 'status' variable, so just remove it.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Otavio Salvador <otavio@ossystems.com.br>
7 years agoinput: Finish simplifing key_matrix_decode_fdt()
Stephen Warren [Thu, 6 Jun 2013 14:48:30 +0000 (10:48 -0400)]
input: Finish simplifing key_matrix_decode_fdt()

[trini: Applied v1 of the series rather than v2, this commit is the
delta from v1 to v2]

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Rini <trini@ti.com>
7 years agoARM: imx: Fix incorrect usage of CONFIG_SYS_MMC_ENV_PART
Fabio Estevam [Tue, 4 Jun 2013 15:05:39 +0000 (15:05 +0000)]
ARM: imx: Fix incorrect usage of CONFIG_SYS_MMC_ENV_PART

When running the "save" command several times on a mx6qsabresd we see:

U-Boot > save
Saving Environment to MMC...
Writing to MMC(1)... done
U-Boot > save
Saving Environment to MMC...
MMC partition switch failed
U-Boot > save
Saving Environment to MMC...
Writing to MMC(1)... done
U-Boot > save
Saving Environment to MMC...
MMC partition switch failed
U-Boot > save
Saving Environment to MMC...
Writing to MMC(1)... done
U-Boot > save
Saving Environment to MMC...
MMC partition switch failed

This issue is caused by the incorrect usage of CONFIG_SYS_MMC_ENV_PART.

CONFIG_SYS_MMC_ENV_PART should be used to specify the mmc partition that stores
the environment variables.

On some imx boards it is been incorrectly used to pass the partition of kernel
and dtb files for the 'mmcpart' script variable.

Remove the CONFIG_SYS_MMC_ENV_PART usage and configure the 'mmcpart' variable
directly.

Reported-by: Jason Liu <r64343@freescale.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Jason Liu <r64343@freescale.com>
7 years agocheckpatch.pl: Add 'printf' to logFunctions
Tom Rini [Thu, 6 Jun 2013 13:28:19 +0000 (09:28 -0400)]
checkpatch.pl: Add 'printf' to logFunctions

Signed-off-by: Tom Rini <trini@ti.com>
7 years agoam33xx/omap4+: Move SRAM_SCRATCH_SPACE_ADDR to <asm/arch/omap.h>
Tom Rini [Thu, 6 Jun 2013 12:57:45 +0000 (08:57 -0400)]
am33xx/omap4+: Move SRAM_SCRATCH_SPACE_ADDR to <asm/arch/omap.h>

The location of valid scratch space is dependent on SoC, so move that
there.  On OMAP4+ we continue to use SRAM_SCRATCH_SPACE_ADDR.  On
am33xx/ti814x we want to use what the ROM defines as "public stack"
which is the area after our defined download image space.  Correct the
comment about and location of CONFIG_SPL_TEXT_BASE.

Signed-off-by: Tom Rini <trini@ti.com>
7 years agoARM: bcm2835: add simplefb DT node during bootz/m
Stephen Warren [Mon, 27 May 2013 18:31:18 +0000 (18:31 +0000)]
ARM: bcm2835: add simplefb DT node during bootz/m

Add a DT simple-framebuffer node to DT when booting the Linux kernel.
This will allow the kernel to inherit the framebuffer configuration from
U-Boot, and display a graphical boot console, and even run a full SW-
rendered X server.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Acked-by: Simon Glass <sjg@chromium.org>
7 years agolcd: add functions to set up simplefb device tree
Stephen Warren [Mon, 27 May 2013 18:31:17 +0000 (18:31 +0000)]
lcd: add functions to set up simplefb device tree

simple-framebuffer is a new device tree binding that describes a pre-
configured frame-buffer memory region and its format. The Linux kernel
contains a driver that supports this binding. Implement functions to
create a DT node (or fill in an existing node) with parameters that
describe the framebuffer format that U-Boot is using.

This will be immediately used by the Raspberry Pi board in U-Boot, and
likely will be used by the Samsung ARM ChromeBook support soon too. It
could well be used by many other boards (e.g. Tegra boards with built-in
LCD panels, which aren't yet supported by the Linux kernel).

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Acked-by: Simon Glass <sjg@chromium.org>
7 years agoMerge branch 'master' of git://git.denx.de/u-boot-arm
Tom Rini [Wed, 5 Jun 2013 16:45:34 +0000 (12:45 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-arm

7 years agoMerge branch 'tpm' of git://git.denx.de/u-boot-x86
Tom Rini [Wed, 5 Jun 2013 12:55:35 +0000 (08:55 -0400)]
Merge branch 'tpm' of git://git.denx.de/u-boot-x86

7 years agoam33xx/omap: Move save_omap_boot_params to omap-common/boot-common.c
Tom Rini [Fri, 31 May 2013 16:31:59 +0000 (12:31 -0400)]
am33xx/omap: Move save_omap_boot_params to omap-common/boot-common.c

We need to call the save_omap_boot_params function on am33xx/ti81xx and
other newer TI SoCs, so move the function to boot-common.  Only OMAP4+
has the omap_hw_init_context function so add ifdefs to not call it on
am33xx/ti81xx.  Call save_omap_boot_params from s_init on am33xx/ti81xx
boards.

Reviewed-by: R Sricharan <r.sricharan@ti.com>
Signed-off-by: Tom Rini <trini@ti.com>
7 years agoam33xx: Correct NON_SECURE_SRAM_START/END
Tom Rini [Fri, 31 May 2013 14:48:03 +0000 (10:48 -0400)]
am33xx: Correct NON_SECURE_SRAM_START/END

Prior to Sricharan's cleanup of the boot parameter saving code, we
did not make use of NON_SECURE_SRAM_START on am33xx, so it wasn't a
problem that the address was pointing to the middle of our running SPL.
Correct to point to the base location of the download image area.
Increase CONFIG_SPL_TEXT_BASE to account for this scratch area being
used.  As part of correcting these tests, make use of the fact that
we've always been placing our stack outside of the download image area
(which is fine, once the downloaded image is run, ROM is gone) so
correct the max size test to be the ROM defined top of the download area
to where we link/load at.

Signed-off-by: Tom Rini <trini@ti.com>
---
Changes in v2:
- Fix typo noted by Peter Korsgaard

7 years agoomap-common/hwinit-common.c: Mark omap_rev_string as static
Tom Rini [Fri, 31 May 2013 14:44:23 +0000 (10:44 -0400)]
omap-common/hwinit-common.c: Mark omap_rev_string as static

Only called in this file, mark as static.

Signed-off-by: Tom Rini <trini@ti.com>
7 years agofdt: allow bootdelay to be specified via device tree
Stephen Warren [Tue, 14 May 2013 08:02:56 +0000 (08:02 +0000)]
fdt: allow bootdelay to be specified via device tree

This can be useful to force bootcmd to execute as soon as U-Boot has
started.

My use-case is: An SoC-specific tool pushes U-Boot into RAM, along with
an image to be written to device boot flash, with the DT config property
"bootcmd" set to contain a command to write that image to flash. In this
scenario, we don't want to allow any stale bootdelay value taken from
the current flash content to affect how long it takes before the
flashing process starts.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
Acked-by: Gerald Van Baren <vanbaren@cideas.com>
7 years agoinput: simplify key_matrix_decode_fdt()
Stephen Warren [Thu, 23 May 2013 12:09:57 +0000 (12:09 +0000)]
input: simplify key_matrix_decode_fdt()

We know the exact property names that the code wants to process. Look
these up directly with fdt_get_property(), rather than iterating over
all properties within the node, and checking each property's name, in
a convoluted fashion, against the expected name.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
7 years agoinput: fix unaligned access in key_matrix_decode_fdt()
Stephen Warren [Wed, 22 May 2013 08:48:18 +0000 (08:48 +0000)]
input: fix unaligned access in key_matrix_decode_fdt()

Initialized character arrays on the stack can cause gcc to emit code that
performs unaligned accessess. Make the data static to avoid this.

Note that the unaligned accesses are made when copying data to prefix[] on
the stack from .rodata. By making the data static, the copy is completely
avoided. All explicitly written code treats the data as u8[], so will never
cause any unaligned accesses.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
7 years agocommon: board_f: Do not call board_postclk_init twice
Masahiro Yamada [Tue, 21 May 2013 21:08:09 +0000 (21:08 +0000)]
common: board_f: Do not call board_postclk_init twice

The generic-board board_init_f function called board_postclk_init twice.

The first one came from arch/arm/lib/board.c, while the second one
from arch/powerpc/lib/board.c.

This commit deletes the first occurrence.
In addition, the second get_clocks call is moved after
board_postclk_init in order to keep the function call order
both for ARM and PowerPC.
ARM board calles get_clocks function after board_postclk_init.

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
7 years agodisk: Fix possible out-of-bounds access in part_efi.c
Marek Vasut [Sun, 19 May 2013 12:53:34 +0000 (12:53 +0000)]
disk: Fix possible out-of-bounds access in part_efi.c

Make sure to never access beyond bounds of either EFI partition name
or DOS partition name. This situation is happening:

part.h:     disk_partition_t->name is 32-byte long
part_efi.h: gpt_entry->partition_name is 36-bytes long

The loop in part_efi.c copies over 36 bytes and thus accesses beyond
the disk_partition_t->name .

Fix this by picking the shortest of source and destination arrays and
make sure the destination array is cleared so the trailing bytes are
zeroed-out and don't cause issues with string manipulation.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Tom Rini <trini@ti.com>
Cc: Simon Glass <sjg@chromium.org>
7 years agosandbox: image: Create a test for loading FIT images
Simon Glass [Thu, 16 May 2013 13:53:28 +0000 (13:53 +0000)]
sandbox: image: Create a test for loading FIT images

The image code is fairly complex with various different options. It would
be useful to have comprehensive tests for this.

As a start, create a script which tries out loading a kernel/ramdisk/fdt
from a FIT and checks that the images appear in the right place in memory.

This uses sandbox which now supports bootm and related features.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agobootstage: Remove unused entries related to kernel/ramdisk/fdt load
Simon Glass [Thu, 16 May 2013 13:53:27 +0000 (13:53 +0000)]
bootstage: Remove unused entries related to kernel/ramdisk/fdt load

Now that the code for loading these three images from a FIT is common, we
don't need individual boostage IDs for each of them.

Note: there are some minor changes in the bootstage numbering, particuarly
for kernel loading. I don't believe this matters.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosandbox: image: Adjust FIT image printing to work with sandbox
Simon Glass [Thu, 16 May 2013 13:53:26 +0000 (13:53 +0000)]
sandbox: image: Adjust FIT image printing to work with sandbox

Use map_sysmem() to convert from address to pointer, so that sandbox can
print FIT information without crashing.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoimage: Use fit_image_load() to load kernel
Simon Glass [Thu, 16 May 2013 13:53:25 +0000 (13:53 +0000)]
image: Use fit_image_load() to load kernel

Use the new common code to load a kernel. The functionality should not
change.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosandbox: Adjust bootm command to work with sandbox
Simon Glass [Thu, 16 May 2013 13:53:24 +0000 (13:53 +0000)]
sandbox: Adjust bootm command to work with sandbox

Use map_sysmem() when converting from addresses to pointers, so that
bootm can be used with sandbox.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoimage: Use fit_image_load() to load FDT
Simon Glass [Thu, 16 May 2013 13:53:23 +0000 (13:53 +0000)]
image: Use fit_image_load() to load FDT

Use the new common code to load a flat device tree. Also fix up a few casts
so that this code works with sandbox. Other than that the functionality
should not change.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoimage: Use fit_image_load() to load ramdisk
Simon Glass [Thu, 16 May 2013 13:53:22 +0000 (13:53 +0000)]
image: Use fit_image_load() to load ramdisk

Use the new common code to load a ramdisk. The functionality should not
change.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoimage: Introduce fit_image_load() to load images from FITs
Simon Glass [Thu, 16 May 2013 13:53:21 +0000 (13:53 +0000)]
image: Introduce fit_image_load() to load images from FITs

At present code to load an image from a FIT is duplicated in the three
places where it is needed (kernel, fdt, ramdisk).

The differences between these different code copies is fairly minor.
Create a new function in the fit code which can handle any of the
requirements of those cases.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agomkimage: Add map_sysmem() and IH_ARCH_DEFAULT to simplfy building
Simon Glass [Thu, 16 May 2013 13:53:20 +0000 (13:53 +0000)]
mkimage: Add map_sysmem() and IH_ARCH_DEFAULT to simplfy building

These are not actually used in mkimage itself, but the image code (which
is common with mkimage) does use them. To avoid #ifdefs in the image code
just for mkimage, define dummy version of these here. The compiler will
eliminate the dead code anyway.

A better way to handle this might be to split out more things from common.h
so that mkimage can include them. At present any file that mkimage uses
has to be very careful what headers it includes.

Signed-off-by: Simon Glass <sjg@chromium.org>