oweals/u-boot.git
9 years agoarc: move SYS_MONITOR_BASE setup in Konfig
Alexey Brodkin [Tue, 3 Feb 2015 10:58:18 +0000 (13:58 +0300)]
arc: move SYS_MONITOR_BASE setup in Konfig

Following SPARK ARC now has SYS_MONITOR_BASE setup via Kconfig.
This makes "include/configs/*.h" cleaner and more flexible.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: hard-code CONFIG_ARCH_EARLY_INIT_R in asm/config.h
Alexey Brodkin [Tue, 3 Feb 2015 10:58:17 +0000 (13:58 +0300)]
arc: hard-code CONFIG_ARCH_EARLY_INIT_R in asm/config.h

Common arch_early_init_r() is used in "arc/lib/cpu.c" for all ARC boards
so there's no sense in separate per-board definitions.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: get rid of useless CONFIG_SKIP_LOWLEVEL_INIT
Alexey Brodkin [Tue, 3 Feb 2015 10:58:16 +0000 (13:58 +0300)]
arc: get rid of useless CONFIG_SKIP_LOWLEVEL_INIT

Currently there's nothing related to really low-level init on ARC so
CONFIG_SKIP_LOWLEVEL_INIT definition makes no sense.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: hard-code CONFIG_SYS_GENERIC_BOARD into asm/config.h
Alexey Brodkin [Tue, 3 Feb 2015 10:58:15 +0000 (13:58 +0300)]
arc: hard-code CONFIG_SYS_GENERIC_BOARD into asm/config.h

There're no other options for ARC except "generic board" so ther's no
point to define CONFIG_SYS_GENERIC_BOARD per board.

We now have it set fo all ARC boards.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: add selection of endianess in Kconfig
Alexey Brodkin [Tue, 3 Feb 2015 10:58:14 +0000 (13:58 +0300)]
arc: add selection of endianess in Kconfig

This change allows to keep board description clean and minimalistic.
This is especially helpful if one board may house different CPUs with
different features.

It is applicable to both FPGA-based boards or those that have CPUs
mounted on interchnagable daughter-boards.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: select cache settings via menuconfig
Alexey Brodkin [Tue, 3 Feb 2015 10:58:13 +0000 (13:58 +0300)]
arc: select cache settings via menuconfig

This change allows to keep board description clean and minimalistic.
This is especially helpful if one board may house different CPUs with
different features.

It is applicable to both FPGA-based boards or those that have CPUs
mounted on interchnagable daughter-boards.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: define and use PTAG AUX regs for MMUv3 only
Alexey Brodkin [Tue, 3 Feb 2015 10:58:12 +0000 (13:58 +0300)]
arc: define and use PTAG AUX regs for MMUv3 only

DC_PTAG and IC_PTAG registers only exist in MMUv3.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: add more flavours of ARC700 series CPU
Alexey Brodkin [Tue, 3 Feb 2015 10:58:11 +0000 (13:58 +0300)]
arc: add more flavours of ARC700 series CPU

Now we may select a particular version of ARC700:
 * ARC750D or
 * ARC770D

It allows more flexible (or more fine tuned) configuration of U-Boot.
Before that change we relied on minimal configuration but now we may
use specific features of each CPU.

Moreover allows us to escape manual selection of options that
exist in both CPUs but may have say different version like MMUv2 in
ARC750D vs MMUv3 in ARC770D.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: remove CPU hard-coded selection from board description in include/configs
Alexey Brodkin [Tue, 3 Feb 2015 10:58:10 +0000 (13:58 +0300)]
arc: remove CPU hard-coded selection from board description in include/configs

With switch to Kconfig we only need very board-specific descriptions in
include/configs.

CPU selection is performed with either defconfig or manually via
menuconfig.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
9 years agoarc: memcmp - fix zero-delay loop utilization
Igor Guryanov [Tue, 3 Feb 2015 10:58:09 +0000 (13:58 +0300)]
arc: memcmp - fix zero-delay loop utilization

It's prohibited to put branch instruction in the very end of zero-delay
loop. On execution this causes "Illegal instruction" exception.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Igor Guryanov <guryanov@synopsys.com>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-ubi
Tom Rini [Wed, 4 Feb 2015 18:30:00 +0000 (13:30 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-ubi

9 years agoPrepare v2015.04-rc1 v2015.04-rc1
Tom Rini [Mon, 2 Feb 2015 17:39:29 +0000 (12:39 -0500)]
Prepare v2015.04-rc1

Signed-off-by: Tom Rini <trini@ti.com>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-ti
Tom Rini [Mon, 2 Feb 2015 17:37:34 +0000 (12:37 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-ti

9 years agoMerge branch 'master' of git://git.denx.de/u-boot-blackfin
Tom Rini [Mon, 2 Feb 2015 16:51:58 +0000 (11:51 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-blackfin

9 years agoMerge branch 'master' of git://git.denx.de/u-boot-sunxi
Tom Rini [Mon, 2 Feb 2015 15:11:44 +0000 (10:11 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-sunxi

9 years agosunxi: mmc: Add 'sunxi_' prefix to the static functions
Siarhei Siamashka [Sat, 31 Jan 2015 22:42:14 +0000 (00:42 +0200)]
sunxi: mmc: Add 'sunxi_' prefix to the static functions

This results in a much more readable callgraph, because now they
can't be confused with the function having exactly the same name
in the generic mmc code.

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: dram: Support more sun[457]i dram parameters in Kconfig
Siarhei Siamashka [Sat, 31 Jan 2015 22:27:06 +0000 (00:27 +0200)]
sunxi: dram: Support more sun[457]i dram parameters in Kconfig

This patch allows to configure all the important DRAM parameters in Kconfig.

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: dram: Optionally use standard JEDEC timings for sun[457]i
Siarhei Siamashka [Sat, 31 Jan 2015 22:27:05 +0000 (00:27 +0200)]
sunxi: dram: Optionally use standard JEDEC timings for sun[457]i

In addition to the current Android magic settings, allow to optionally use
DDR3 timing parameters, which are tailored for different clock frequencies
and JEDEC speed bins. This should improve reliability and performance.

Adding '+S:CONFIG_DRAM_TIMINGS_DDR3_1066F_1333H=y' to the board defconfig
allows to use timings, which are calculated for the DDR3-1066F speed bin.
A lot of DDR3 chips, which are used in real Allwinner based devices,
support DDR3-1066F speed bin timings.

And adding '+S:CONFIG_DRAM_TIMINGS_DDR3_800E_1066G_1333J=y' should work
with any DDR3 chips, because this targets the slowest JEDEC speed bins.

The vendor magic values are still used by default for DRAM, but board
maintainers now have more flexibility in DRAM timings selection.

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: video: Force h/vsync active high when using ext. vga dac on some boards
Hans de Goede [Sun, 25 Jan 2015 14:33:07 +0000 (15:33 +0100)]
sunxi: video: Force h/vsync active high when using ext. vga dac on some boards

On both my A13-OLinuxIno and my A13-OLinuxIno-Micro, the vga output gives an
unstable image when active low v or hsync is used.

The problem seems to be specific to the OLinuxIno A13 (normal & micro)
boards. I've just looked up the schematics and they use an opendrain driver
for the vga sync lines, and with sync pulses it is the logical high->low
edge of the pulse which counts for the timing, which with an active low
sync is being driven by the pull-up, and that simply seems to not drive
it hard enough to get a stable image.

So force v and hsync active high on these boards. independent of what the
modeline says. This fixes the unstable image.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: Add Inet 86VS support
Michal Suchanek [Sun, 4 Jan 2015 20:06:21 +0000 (21:06 +0100)]
sunxi: Add Inet 86VS support

Signed-off-by: Michal Suchanek <hramrach@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: TZX-Q8-713B7 support
Paul Kocialkowski [Tue, 27 Jan 2015 18:58:46 +0000 (19:58 +0100)]
sunxi: TZX-Q8-713B7 support

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: Add Linksprite_pcDuino3_Nano board / defconfig
Adam Sampson [Sun, 1 Feb 2015 12:58:58 +0000 (12:58 +0000)]
sunxi: Add Linksprite_pcDuino3_Nano board / defconfig

This is a low-cost Allwinner A20 board with Arduino-style GPIO headers;
it features 1G RAM, 4G NAND flash, 1 micro-SD, 2 USB sockets, 1 micro
USB socket for OTG and another for power in, HDMI, SATA, 5V power for
SATA devices, gigabit Ethernet, an IR receiver, 3.5mm audio out and a
MIPI camera connector.

Like the BananaPi, this board needs GMAC_TX_DELAY set to 3 in order for
GMAC to work reliably at gigabit speeds.

For more details, see: http://linux-sunxi.org/LinkSprite_pcDuino3_Nano

Signed-off-by: Adam Sampson <ats@offog.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: MAINTAINERS: sort entries alphabetically
Hans de Goede [Wed, 28 Jan 2015 08:31:52 +0000 (09:31 +0100)]
sunxi: MAINTAINERS: sort entries alphabetically

Keep all entries except for the monster entry at the top alphabetically sorted.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: rsb: Move rsb_set_device_mode() call to rsb_init()
Hans de Goede [Mon, 26 Jan 2015 15:59:12 +0000 (16:59 +0100)]
sunxi: rsb: Move rsb_set_device_mode() call to rsb_init()

It turns out that the device_mode_data is rsb specific, rather then slave
specific, so integrate the rsb_set_device_mode() call into rsb_init().

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: rsb: Add sun9i (A80 support)
Hans de Goede [Mon, 26 Jan 2015 15:46:43 +0000 (16:46 +0100)]
sunxi: rsb: Add sun9i (A80 support)

Add support for the A80 to the rsb code.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: Remove CONFIG_TARGET_FOO for sun5i and sun7i boards
Hans de Goede [Sun, 25 Jan 2015 11:15:14 +0000 (12:15 +0100)]
sunxi: Remove CONFIG_TARGET_FOO for sun5i and sun7i boards

CONFIG_TARGET_FOO was only used in board/sunxi/Makefile to select the
dram config for sun5i and sun7i boards and in board/sunxi/gmac.c for some
special handling of the bananapi/bananapro (both sun7i), all sun5i and sun7i
boards have been moved over to using a single dram_sun5i_autoconfig file,
and the tx clk delay handling for the Banana boards now has its own Kconfig.

IOW nothing is using CONFIG_TARGET_FOO anymore, so remove it.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: Add a GMAC Transmit Clock Delay Chain Kconfig option
Hans de Goede [Sun, 25 Jan 2015 11:10:48 +0000 (12:10 +0100)]
sunxi: Add a GMAC Transmit Clock Delay Chain Kconfig option

And use this to set the GMAC Transmit Clock Delay Chain value on Banana
boards, rather then keying of CONFIG_TARGET_FOO.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: Convert sun5i boards to use auto dram configuration
Hans de Goede [Sun, 25 Jan 2015 10:29:27 +0000 (11:29 +0100)]
sunxi: Convert sun5i boards to use auto dram configuration

Currently we've separate detailed dram settings for all sun5i boards, this
moves them over to using auto dram configuration so that we can get rid of
all the per board dram_foo.c files.

This has been tested on a A10s-Olinuxino, A13-Olinuxino, A13-OlinuxinoM,
mk802-a10s and r7-tv-dongle board.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: Hyundai_A7HD_defconfig fix USB vbus pin config
Hans de Goede [Fri, 23 Jan 2015 15:40:01 +0000 (16:40 +0100)]
sunxi: Hyundai_A7HD_defconfig fix USB vbus pin config

USB1_VBUS is not used, and USB2_VBUS uses the pin normally used to control
USB1_VBUS.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: Only enable i2c support in the SPL when needed
Hans de Goede [Fri, 23 Jan 2015 14:28:22 +0000 (15:28 +0100)]
sunxi: Only enable i2c support in the SPL when needed

We do not need i2c support in the SPL when there is no PMIC (some sun4i
boards), or when the PMIC is not using i2c such as on sun6i and sun8i.

This reduces the SPL size from (e.g.) 21812 to 19260 bytes.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agoMerge branch 'patman' of git://git.denx.de/u-boot-x86
Tom Rini [Sat, 31 Jan 2015 17:40:48 +0000 (12:40 -0500)]
Merge branch 'patman' of git://git.denx.de/u-boot-x86

9 years agoMerge branch 'master' of git://git.denx.de/u-boot-net
Tom Rini [Sat, 31 Jan 2015 17:40:26 +0000 (12:40 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-net

9 years agopatman: Explain how to make doc/git-mailrc work
Simon Glass [Sat, 4 Oct 2014 02:40:36 +0000 (20:40 -0600)]
patman: Explain how to make doc/git-mailrc work

Add an explanation for how to set up git so that patman can find the alias
file. Fix up the get_maintainers message too.

Reported-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Fix README to indicate that vendor name is unset
Simon Glass [Tue, 23 Sep 2014 19:05:59 +0000 (13:05 -0600)]
sandbox: Fix README to indicate that vendor name is unset

This brings in a additional small fix which was missed in a recent update
to the README.

Suggested-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
9 years agopatman: Check commit_match before stripping leading whitespace
Scott Wood [Thu, 25 Sep 2014 19:30:46 +0000 (14:30 -0500)]
patman: Check commit_match before stripping leading whitespace

True commit lines start at column zero.  Anything that is indented
is part of the commit message instead.  I noticed this by trying to
run buildman with commit e3a4facdfc07179ebe017a07b8de6224a935a9f3
as master, which contained a reference to a Linux commit inside
the commit message.  ProcessLine saw that as a genuite commit
line, and thus buildman tried to build it, and died with an
exception because that SHA is not present in the U-Boot tree.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agopatman: Make dry-run output match real functionality
Peter Tyser [Mon, 26 Jan 2015 17:42:21 +0000 (11:42 -0600)]
patman: Make dry-run output match real functionality

When run with the --dry-run argument patman prints out information
showing what it would do.  This information currently doesn't line up
with what patman/git send-email really do.  Some basic examples:
- If an email address is addressed via "Series-cc" and "Patch-cc" patman
  shows that email address would be CC-ed two times.
- If an email address is addressed via "Series-to" and "Patch-cc" patman
  shows that email address would be sent TO and CC-ed.
- If an email address is addressed from a combination of tag aliases,
  get_maintainer.pl output, "Series-cc", "Patch-cc", etc patman shows
  that the email address would be CC-ed multiple times.

Patman currently does try to send duplicate emails like the --dry-run
output shows, but "git send-email" intelligently removes duplicate
addresses so this patch shouldn't change the non-dry-run functionality.

Change patman's output and email addressing to line up with the
"git send-email" logic.  This trims down patman's dry-run output and
prevents confusion about what patman will do when emails are actually
sent.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Add a -D option to use a default device tree
Simon Glass [Tue, 20 Jan 2015 03:21:34 +0000 (20:21 -0700)]
sandbox: Add a -D option to use a default device tree

It is painful to specify the full path to the device tree with the -d
option. It is normally kept in the same directory as U-Boot, so provide
an option to use this by default.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Correct cros-ec keyboard definition
Simon Glass [Tue, 20 Jan 2015 03:21:33 +0000 (20:21 -0700)]
sandbox: Correct cros-ec keyboard definition

The other boards got updated to the standard binding. Update sandbox as
well.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agonet/designware: add error message on DMA reset timeout
Alexey Brodkin [Tue, 13 Jan 2015 14:10:24 +0000 (17:10 +0300)]
net/designware: add error message on DMA reset timeout

If for some reason DMA module fails to reset user oserves only this:
--->---
# dhcp
Trying dwmac.e0018000
FAIL
--->---

This message makes not much sense.
With proposed change error message will be more helpful:
--->---
# dhcp
Trying dwmac.e0018000
DMA reset timeout
FAIL
--->---

For example user may do power toggle to recover board functionality.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Chin Liang See <clsee@altera.com>
Cc: Dinh Nguyen <dinguyen@altera.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Pavel Machek <pavel@denx.de>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
Cc: Marek Vasut <marex@denx.de>
Reviewed-by: Tom Rini <trini@ti.com>
Acked-by: Pavel Machek <pavel@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agonet: phy: micrel: add support for KSZ8895 switch in SMI mode
Philippe De Muyter [Wed, 19 Feb 2014 16:21:59 +0000 (17:21 +0100)]
net: phy: micrel: add support for KSZ8895 switch in SMI mode

This patch adds a phy driver for the Micrel KSZ8895 switch.  As the SoC MAC
is directly connected to the switch MAC the link to the switch is always up.

But the KSZ8895 switch can be hardwired in three configuration modes :
- not configurable with eventually an eeprom-stored configuration
- configurable by the mdio/mdc connection (SMI protocol)
- configurable by a SPI connection.

In not configurable mode, the switch starts automatically, but in the
other modes, it must be started programmatically, by writing 1 in
configuration register 1.
We only support the not configurable and mdio/mdc (aka SMI) modes here.

Signed-off-by: Philippe De Muyter <phdm@macqel.be>
Cc: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agoAdd MS7206SE ethernet support
Yoshinori Sato [Thu, 13 Feb 2014 14:13:41 +0000 (23:13 +0900)]
Add MS7206SE ethernet support

Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agonet: tsec: Fix NULL access in case init_phy() fails
Claudiu Manoil [Tue, 10 Dec 2013 13:21:04 +0000 (15:21 +0200)]
net: tsec: Fix NULL access in case init_phy() fails

If the PHY is not recognized don't access phydev (NULL)
and return 0 to signal failure.

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-mips
Tom Rini [Fri, 30 Jan 2015 18:56:15 +0000 (13:56 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-mips

9 years agoMerge branch 'master' of git://git.denx.de/u-boot-dm
Tom Rini [Fri, 30 Jan 2015 14:24:42 +0000 (09:24 -0500)]
Merge branch 'master' of git://git.denx.de/u-boot-dm

9 years agoARM: armv7 fix spelling of SCTRL
Peng Fan [Thu, 29 Jan 2015 10:03:39 +0000 (18:03 +0800)]
ARM: armv7 fix spelling of SCTRL

SCTLR is the abbreviation of System Control Register, so we should
use SCTLR but not SCTRL.

Signed-off-by: Peng Fan <Peng.Fan@freescale.com>
9 years agofpga: Extend dump description
Michal Simek [Mon, 26 Jan 2015 07:52:27 +0000 (08:52 +0100)]
fpga: Extend dump description

There are missing parameters in help which fpga dump command
requires.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
9 years agovexpress64: support the Juno Development Platform
Linus Walleij [Fri, 23 Jan 2015 13:41:10 +0000 (14:41 +0100)]
vexpress64: support the Juno Development Platform

The Juno Development Platform is a physical Versatile Express
device with some differences from the emulated semihosting
models. The main difference is that the system is split in
a SoC and an FPGA where the SoC hosts the serial ports at
totally different adresses.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
9 years agovexpress64: get rid of CONFIG_SYS_EXTRA_OPTIONS
Linus Walleij [Fri, 23 Jan 2015 10:50:53 +0000 (11:50 +0100)]
vexpress64: get rid of CONFIG_SYS_EXTRA_OPTIONS

The Versatile Express ARMv8 semihosted FVP platform is still
using the legacy CONFIG_SYS_EXTRA_OPTIONS method to configure
some compile-time flags. Get rid of this and create a Kconfig
entry for the FVP model, and a selectable bool for the
semihosting library.

The FVP subboard is now modeled as a target choice so we can
eventually choose between different ARMv8 versatile express
boards (FVP, base model, Juno...) this way. All dependent
symbols are updated to reflect this.

The 64bit Versatile Express board symbols are renamed
VEXPRESS64 so we have some chance to see what is actually
going on. Tested on the FVP fast model.

Acked-by: Steve Rae <srae@broadcom.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
9 years agoadd README.distro file
Dennis Gilmore [Thu, 22 Jan 2015 18:34:20 +0000 (11:34 -0700)]
add README.distro file

Add documentation on how to setup a system to use the generic distro
configs and boot commands. This spells out what is needed to make a
system conformant, but does not limit the board to only the defaults.

Signed-off-by: Dennis Gilmore <dennis@ausil.us>
[swarren, added concept, user config, BOOT_TARGET_DEVICES sections.
edited the rest]
Signed-off-by: Stephen Warren <swarren@nvidia.com>
9 years agodrivers/pci/pci_rom.c: fix compile warning under 64bit mode
Minghuan Lian [Thu, 22 Jan 2015 05:21:55 +0000 (13:21 +0800)]
drivers/pci/pci_rom.c: fix compile warning under 64bit mode

Fix this:
drivers/pci/pci_rom.c:95:15: warning: cast to pointer from
integer of different size [-Wint-to-pointer-cast]
rom_header = (struct pci_rom_header *)rom_address;

Signed-off-by: Minghuan Lian <Minghuan.Lian@freescale.com>
9 years agodrivers/net/e1000.c: fix compile warning under 64bit mode
Minghuan Lian [Thu, 22 Jan 2015 05:21:54 +0000 (13:21 +0800)]
drivers/net/e1000.c: fix compile warning under 64bit mode

Fix this:
warning: cast from pointer to integer of different size

Signed-off-by: Minghuan Lian <Minghuan.Lian@freescale.com>
9 years agoarm: switch armltd vexpress to GENERIC_BOARD
Chris Kuethe [Thu, 22 Jan 2015 04:52:33 +0000 (20:52 -0800)]
arm: switch armltd vexpress to GENERIC_BOARD

only tested tested under QEMU with vexpress_ca9x4 ("-M vexpress-a9") and
vexpress_ca15_tc2 ("-M vexpress-a15"). Makes the ugly warning go away.

Signed-off-by: Chris Kuethe <chris.kuethe+github@gmail.com>
9 years agodistro_bootcmd: read DHCP boot script name from a variable
Stephen Warren [Mon, 19 Jan 2015 23:39:11 +0000 (16:39 -0700)]
distro_bootcmd: read DHCP boot script name from a variable

Modify $bootcmd_dhcp to read the downloaded script filename from an
environment variable rather than hard-coding it. This allows the user
(or another script) to select a different script name if they want,
without editing the whole value of $bootcmd_dhcp.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
9 years agocmd: Fix gettime command help
Bin Meng [Mon, 19 Jan 2015 13:32:00 +0000 (21:32 +0800)]
cmd: Fix gettime command help

Remove the additional ',' and '\n' from the gettime command help.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoMIPS: unify CPU code in arch/mips/cpu/
Daniel Schwierzeck [Thu, 29 Jan 2015 13:56:20 +0000 (14:56 +0100)]
MIPS: unify CPU code in arch/mips/cpu/

Unify and move code in arch/mips/cpu/mips[32|64]/ to arch/mips/cpu/.
The CPU specific config.mk files need to remain until
CONFIG_STANDALONE_LOAD_ADDR is converted to a global Kconfig symbol.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
9 years agoMIPS: move au1x00 SoC code to arch/mips/mach-au1x00
Daniel Schwierzeck [Thu, 29 Jan 2015 13:47:01 +0000 (14:47 +0100)]
MIPS: move au1x00 SoC code to arch/mips/mach-au1x00

Move all au1x00 code out of arch/mips/cpu/mips32 to allow
unification of CPU code in a later patch. The reorganization
of the SoC specific header files will be done in a later patch
series.

Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Cc: Paul Burton <paul.burton@imgtec.com>
9 years agoMIPS: handle mips64 ST0_KX bit in mips32 start.S
Paul Burton [Thu, 29 Jan 2015 10:04:10 +0000 (10:04 +0000)]
MIPS: handle mips64 ST0_KX bit in mips32 start.S

In preparation for sharing a single copy of start.S between mips32 &
mips64, handle setting the KX bit of the cop0 Status register when the
mips32 start.S is built for mips64.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
9 years agoMIPS: handle mips64 relocs in mips32 start.S
Paul Burton [Thu, 29 Jan 2015 10:04:09 +0000 (10:04 +0000)]
MIPS: handle mips64 relocs in mips32 start.S

In preparation for sharing a single copy of start.S between mips32 &
mips64, handle mips64 relocations in the mips32 start.S when built for
mips64.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
9 years agoMIPS: use asm.h macros in mips32 start.S
Paul Burton [Thu, 29 Jan 2015 10:04:08 +0000 (10:04 +0000)]
MIPS: use asm.h macros in mips32 start.S

Where the mips32 & mips64 implementations of start.S differ in terms of
access sizes & offsets, use the appropriate macros from asm.h to
abstract those differences away. This is in preparation for sharing a
single copy of start.S between mips32 & mips64.

The exception to this is loads of immediates to be written to the cop0
Config register, which is a 32bit register on mips64 and therefore
constants written to it can be loaded as such.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
9 years agodm: cros_ec: Convert cros_ec_i2c over to driver model
Simon Glass [Tue, 27 Jan 2015 03:29:40 +0000 (20:29 -0700)]
dm: cros_ec: Convert cros_ec_i2c over to driver model

Move this driver to use driver model and update the snow configuration to
match.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: i2c: Add two more I2C init functions to the compatibility layer
Simon Glass [Tue, 27 Jan 2015 03:29:39 +0000 (20:29 -0700)]
dm: i2c: Add two more I2C init functions to the compatibility layer

These functions are useful in case the board calls them. Also fix a missing
parameter caused by applying the wrong patch (actually I failed to send v2
and applied v1 by mistake).

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: exynos: dts: Set the offset length for cros_ec
Simon Glass [Tue, 27 Jan 2015 03:29:38 +0000 (20:29 -0700)]
dm: exynos: dts: Set the offset length for cros_ec

The EC has no concept of offset, so use a value of 0.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: i2c: dts: Support an offset-len device tree property
Simon Glass [Tue, 27 Jan 2015 03:29:37 +0000 (20:29 -0700)]
dm: i2c: dts: Support an offset-len device tree property

Since U-Boot can support different offset lengths (0-4 bytes), add a device
tree property to specify this. This avoids hard-coding it in the driver.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoexynos5: enable dm i2c
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:39 +0000 (13:36 +0100)]
exynos5: enable dm i2c

This patch enables CONFIG_DM_I2C and also CONFIG_DM_I2C_COMPAT.
The last one should be removed when all the i2c peripheral
drivers will use dm i2c framework.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
9 years agoodroid u3: enable dm i2c support
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:38 +0000 (13:36 +0100)]
odroid u3: enable dm i2c support

This patch enables CONFIG_DM_I2C and also CONFIG_DM_I2C_COMPAT.
The last one should be removed when the dm pmic framework will
be finished.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
9 years agoodroid u3: dts: add missing i2c aliases
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:37 +0000 (13:36 +0100)]
odroid u3: dts: add missing i2c aliases

This change fixes i2c bus numbering for Odroid U3.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
9 years agodm: i2c: s3c24x0: adjust to dm-i2c api
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:36 +0000 (13:36 +0100)]
dm: i2c: s3c24x0: adjust to dm-i2c api

This commit adjusts the s3c24x0 driver to new i2c api
based on driver-model. The driver supports standard
and high-speed i2c as previous.

Tested on Trats2, Odroid U3, Arndale, Odroid XU3

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Tested-by: Simon Glass <sjg@chromium.org>
Cc: Simon Glass <sjg@chromium.org>
Cc: Heiko Schocher <hs@denx.de>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoexynos5: pinmux: check flag for i2c config
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:34 +0000 (13:36 +0100)]
exynos5: pinmux: check flag for i2c config

Some versions of Exynos5 supports High-Speed I2C,
on few interfaces, this change allows support this.
The new flag is: PINMUX_FLAG_HS_MODE

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoarndale: dts: add missing i2c aliases
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:33 +0000 (13:36 +0100)]
arndale: dts: add missing i2c aliases

Without this alias setting, the seq numbers
of the i2c devices are wrong.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
9 years agoexynos4: dts: add missing i2c properties
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:32 +0000 (13:36 +0100)]
exynos4: dts: add missing i2c properties

This patch modify i2c nodes in exynos4.dtsi with:
- adding proper interrupts arrays for each i2c node,
  which allows to decode periph id
- add reg address for each i2c node for i2c driver internal use

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Lukasz Majewski <l.majewski@samsung.com>
9 years agosmdk5250: config: enable max77686 driver support
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:31 +0000 (13:36 +0100)]
smdk5250: config: enable max77686 driver support

This commit enable support for the above driver,
which was disabled in common config.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
9 years agoexynos5250: config: disable max77686 driver
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:30 +0000 (13:36 +0100)]
exynos5250: config: disable max77686 driver

This PMIC is not common for all Exynos5250
based boards, so should be romoved from
common config.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Cc: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
9 years agoarndale: config: disable max77686 support
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:29 +0000 (13:36 +0100)]
arndale: config: disable max77686 support

There is no MAX77686 pmic on this board,
so the driver support should be removed.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
9 years agodm: i2c-uclass-compat: fix missed argument
Przemyslaw Marczak [Tue, 27 Jan 2015 12:36:28 +0000 (13:36 +0100)]
dm: i2c-uclass-compat: fix missed argument

This patch fixes build error for CONFIG_DM_I2C_COMPAT.
In i2c_get_chip_for_busnum() call, one of argument was missed,
which was offset_len. Now it is set to 'alen' as previous.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agorsa: Use checksum algorithms from struct hash_algo
Ruchika Gupta [Fri, 23 Jan 2015 10:31:59 +0000 (16:01 +0530)]
rsa: Use checksum algorithms from struct hash_algo

Currently the hash functions used in RSA are called directly from the sha1
and sha256 libraries. Change the RSA checksum library to use the progressive
hash API's registered with struct hash_algo. This will allow the checksum
library to use the hardware accelerated progressive hash API's once available.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
(Fixed build error in am335x_boneblack_vboot due to duplicate CONFIG_DM)

Change-Id: Ic44279432f88d4e8594c6e94feb1cfcae2443a54

9 years agoUse hash.c in mkimage
Ruchika Gupta [Fri, 23 Jan 2015 10:31:58 +0000 (16:01 +0530)]
Use hash.c in mkimage

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agohash: Add function to find hash_algo struct with progressive hash
Ruchika Gupta [Fri, 23 Jan 2015 10:31:57 +0000 (16:01 +0530)]
hash: Add function to find hash_algo struct with progressive hash

The hash_algo structure has some implementations in which progressive hash
API's are not defined. These are basically the hardware based implementations
of SHA. An API is added to find the algo which has progressive hash API's
defined. This can then be integrated with RSA checksum library which uses
Progressive Hash API's.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agolib/rsa: Add Kconfig for devices supporting RSA Modular Exponentiation
Ruchika Gupta [Fri, 23 Jan 2015 10:31:56 +0000 (16:01 +0530)]
lib/rsa: Add Kconfig for devices supporting RSA Modular Exponentiation

Kconfig option added for devices which support RSA Verification.
1. RSA_SOFTWARE_EXP
Enables driver for supporting RSA Modular Exponentiation in Software
2. RSA_FREESCALE_EXP
Enables driver for supporting RSA Modular Exponentiation using Freescale specific
driver

The above drivers use RSA uclass

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
(Removed duplicate line in Kconfig comment)

Change-Id: I7663c4d5350e2bfc3dfa2696f70ef777d6ccc6f6

9 years agoDM: crypto/fsl - Add Freescale rsa DM driver
Ruchika Gupta [Fri, 23 Jan 2015 10:31:55 +0000 (16:01 +0530)]
DM: crypto/fsl - Add Freescale rsa DM driver

Driver added for RSA Modular Exponentiation using Freescale Hardware
Accelerator CAAM. The driver uses UCLASS_MOD_EXP

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agolib/rsa: Modify rsa to use DM driver
Ruchika Gupta [Fri, 23 Jan 2015 10:31:54 +0000 (16:01 +0530)]
lib/rsa: Modify rsa to use DM driver

Modify rsa_verify to use the rsa driver of DM library .The tools
will continue to use the same RSA sw library.

CONFIG_RSA is now dependent on CONFIG_DM. All configurations which
enable FIT based signatures have been modified to enable CONFIG_DM
by default.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoconfigs: Move CONFIG_FIT_SIGNATURE to defconfig
Ruchika Gupta [Fri, 23 Jan 2015 10:31:53 +0000 (16:01 +0530)]
configs: Move CONFIG_FIT_SIGNATURE to defconfig

For the platforms which use,CONFIG_FIT_SIGNATURE, the required configs are
moved to the platform's defconfig file. Selecting CONFIG_FIT_SIGNATURE using
defconfig automatically resolves the dependencies for signature verification.
The RSA library gets automatically selected and user does not have to define
CONFIG_RSA manually.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoDM: crypto/rsa_mod_exp: Add rsa Modular Exponentiation DM driver
Ruchika Gupta [Fri, 23 Jan 2015 10:31:52 +0000 (16:01 +0530)]
DM: crypto/rsa_mod_exp: Add rsa Modular Exponentiation DM driver

Add a new rsa uclass for performing modular exponentiation and implement
the software driver basing on this uclass.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoFIT: Modify option FIT_SIGNATURE in Kconfig
Ruchika Gupta [Fri, 23 Jan 2015 10:31:51 +0000 (16:01 +0530)]
FIT: Modify option FIT_SIGNATURE in Kconfig

For FIT signature based approach to work, RSA library needs to be selected.
The FIT_SIGNATURE option in Kconfig is modified to automatically select RSA.
Selecting RSA compiles the RSA library required for image verification.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agorsa: Split the rsa-verify to separate the modular exponentiation
Ruchika Gupta [Fri, 23 Jan 2015 10:31:50 +0000 (16:01 +0530)]
rsa: Split the rsa-verify to separate the modular exponentiation

Public exponentiation which is required in rsa verify functionality is
tightly integrated with verification code in rsa_verify.c. The patch
splits the file into twp separating the modular exponentiation.

1. rsa-verify.c
- The file parses device tree keys node to fill a keyprop structure.
The keyprop structure can then be converted to implementation specific
format.
(struct rsa_pub_key for sw implementation)
- The parsed device tree node is then passed to a generic rsa_mod_exp
function.

2. rsa-mod-exp.c
Move the software specific functions related to modular exponentiation
from rsa-verify.c to this file.

Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
CC: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoExport redesign
Martin Dorwig [Mon, 26 Jan 2015 22:22:54 +0000 (15:22 -0700)]
Export redesign

this is an atempt to make the export of functions typesafe.
I replaced the jumptable void ** by a struct (jt_funcs) with function pointers.
The EXPORT_FUNC macro now has 3 fixed parameters and one
variadic parameter
The first is the name of the exported function,
the rest of the parameters are used to format a functionpointer
in the jumptable,

the EXPORT_FUNC macros are expanded three times,
1. to declare the members of the struct
2. to initialize the structmember pointers
3. to call the functions in stubs.c

Signed-off-by: Martin Dorwig <dorwig@tetronik.com>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
(resending to the list since my tweaks are not quite trivial)

9 years agodm: Update documentation for new bus features
Simon Glass [Sun, 25 Jan 2015 15:27:20 +0000 (08:27 -0700)]
dm: Update documentation for new bus features

Now that we have new bus features, update README.txt and the SPI docs to
explain these.
Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: cros_ec_spi: Remove old pre-driver-model code
Simon Glass [Sun, 25 Jan 2015 15:27:19 +0000 (08:27 -0700)]
dm: cros_ec_spi: Remove old pre-driver-model code

This is no-longer needed since all platforms use SPI for cros_ec.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: cros_ec: Don't require protocol 3 support
Simon Glass [Sun, 25 Jan 2015 15:27:17 +0000 (08:27 -0700)]
dm: cros_ec: Don't require protocol 3 support

I2C is now deprecated on ARM platforms and there are no devices that use it
with the v3 protocol. We can't require v3 support if we want to support I2C.
Adjust the error handling to suit.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Ignore disabled devices when binding
Simon Glass [Sun, 25 Jan 2015 15:27:16 +0000 (08:27 -0700)]
dm: core: Ignore disabled devices when binding

We don't want to bind devices which should never be used.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
9 years agodm: exynos: Drop unused COMPAT features for SPI
Simon Glass [Sun, 25 Jan 2015 15:27:15 +0000 (08:27 -0700)]
dm: exynos: Drop unused COMPAT features for SPI

This has moved to driver model so we don't need the fdtdec support.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Minkyu Kang <mk7.kang@samsung.com>
9 years agodm: tegra: Drop unused COMPAT features for I2C, SPI
Simon Glass [Sun, 25 Jan 2015 15:27:14 +0000 (08:27 -0700)]
dm: tegra: Drop unused COMPAT features for I2C, SPI

These have moved to driver model so we don't need the fdtdec support.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: i2c: Move slave details to child platdata
Simon Glass [Sun, 25 Jan 2015 15:27:13 +0000 (08:27 -0700)]
dm: i2c: Move slave details to child platdata

At present we go through various contortions to store the I2C's chip
address in its private data. This only exists when the chip is active so
must be set up when it is probed. Until the device is probed we don't
actually record what address it will appear on.

However, now that we can support per-child platform data, we can use that
instead. This allows us to set up the address when the child is bound,
and avoid the messy contortions.

Unfortunately this is a fairly large change and it seems to be difficult to
break it down further.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
9 years agodm: spi: Move slave details to child platdata
Simon Glass [Sun, 25 Jan 2015 15:27:12 +0000 (08:27 -0700)]
dm: spi: Move slave details to child platdata

At present we go through various contortions to store the SPI slave's chip
select in its private data. This only exists when the slave is active so
must be set up when it is probed. Until the device is probed we don't
actually know what chip select it will appear on.

However, now that we can support per-child platform data, we can use that
instead. This allows us to set up the chip select when the child is bound,
and avoid the messy contortions.

Unfortunately this is a fairly large change and it seems to be difficult to
break it down further.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: spi: Set up the spi_slave device pointer in child_pre_probe()
Simon Glass [Sun, 25 Jan 2015 15:27:11 +0000 (08:27 -0700)]
dm: spi: Set up the spi_slave device pointer in child_pre_probe()

At present we use struct spi_slave as our device pointer in a lot of places
to avoid changing the old SPI API. At some point this will go away.

But for now, it is better if the SPI uclass sets up this pointer, rather
than relying on passing it into the device when it is probed. We can use the
new uclass child_pre_probe() method to do this.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Allow uclass to set up a device's child before it is probed
Simon Glass [Sun, 25 Jan 2015 15:27:10 +0000 (08:27 -0700)]
dm: core: Allow uclass to set up a device's child before it is probed

Some buses need to set up their devices before they can be used. This setup
may well be common to all buses in a particular uclass. Support a common
pre-probe method for the uclass, called before any bus devices are probed.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
9 years agodm: sandbox: sf: Tidy up the error handling in sandbox_sf_probe()
Simon Glass [Sun, 25 Jan 2015 15:27:09 +0000 (08:27 -0700)]
dm: sandbox: sf: Tidy up the error handling in sandbox_sf_probe()

Use a single exit point when we have an error and add debugging there.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Allow the uclass to set up a device's child after binding
Simon Glass [Sun, 25 Jan 2015 15:27:08 +0000 (08:27 -0700)]
dm: core: Allow the uclass to set up a device's child after binding

For buses, after a child is bound, allow the uclass to perform some
processing. This can be used to figure out the address of the child (e.g.
the chip select for SPI slaves) so that it is ready to be probed.

This avoids bus drivers having to repeat the same process, which really
should be done by the uclass, since it is common.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
9 years agodm: spi: Move the per-child data size to the uclass
Simon Glass [Sun, 25 Jan 2015 15:27:07 +0000 (08:27 -0700)]
dm: spi: Move the per-child data size to the uclass

This is common to all SPI drivers and specifies a structure used by the
uclass. It makes more sense to define it in the uclass.

Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Allow uclasses to specify private data for a device's children
Simon Glass [Sun, 25 Jan 2015 15:27:06 +0000 (08:27 -0700)]
dm: core: Allow uclasses to specify private data for a device's children

In many cases the per-child private data for a device's children is defined
by the uclass rather than the individual driver. For example, a SPI bus
needs to store information about each of its children, but all SPI drivers
store the same information. It makes sense to allow the uclass to define
this data.

If the driver provides a size value for its per-child private data, then use
it. Failng that, fall back to that provided by the uclass.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
9 years agodm: core: Add a flag to control sequence numbering
Simon Glass [Sun, 25 Jan 2015 15:27:05 +0000 (08:27 -0700)]
dm: core: Add a flag to control sequence numbering

At present we try to use the 'reg' property and device tree aliases to give
devices a sequence number. The 'reg' property is often actually a memory
address, so the sequence numbers thus-obtained are not useful. It would be
better if the devices were just sequentially numbered in that case. In fact
neither I2C nor SPI use this feature, so drop it.

Some devices need us to look up an alias to number them within the uclass.
Add a flag to control this, so it is not done unless it is needed.

Adjust the tests to test this new behaviour.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>