project/bcm63xx/u-boot.git
13 years agocmd_bdinfo: replace print_str() with print_mhz()
Timur Tabi [Wed, 5 Oct 2011 22:08:07 +0000 (17:08 -0500)]
cmd_bdinfo: replace print_str() with print_mhz()

The print_str() helper function for cmd_bdinfo can print any string, but it
is only used to print MHz values.  Replace it with print_mhz() that takes
a number and converts it to a string internally.

Signed-off-by: Timur Tabi <timur@freescale.com>
13 years agogpio: Add PCA9698 40-bit I2C I/O port
Dirk Eibach [Mon, 3 Oct 2011 23:13:51 +0000 (23:13 +0000)]
gpio: Add PCA9698 40-bit I2C I/O port

Signed-off-by: Dirk Eibach <eibach@gdsys.de>
13 years agoserial: pl01x: drain PL01x FIFO before baudrate change
Linus Walleij [Sun, 2 Oct 2011 11:52:52 +0000 (11:52 +0000)]
serial: pl01x: drain PL01x FIFO before baudrate change

Not draining the FIFO and waiting for the UART to be non-busy
before changing baudrate results in crap characters on the
console, so let's wait for the FIFO to drain and the last
character to be clocked out before we do that.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
13 years agoexamples: api: allow build with private libgcc
Che-liang Chiou [Mon, 17 Oct 2011 21:04:15 +0000 (21:04 +0000)]
examples: api: allow build with private libgcc

The examples/api is not configured with USE_PRIVATE_LIBGCC.  This makes
building examples/api break on certain boards that do not/cannot use the
public libgcc.

Nevertheless, this patch has to also touch the top-level Makefile to fix
this problem because the current top-level Makefile does not specify
libgcc as a prerequisite of examples/api, and explicitly builds
examples/api _before_ libgcc.

For testing this patch, I added the following to configs/seaboard.h and
ran demo.bin on a Seaboard.

+#define CONFIG_API
+#define CONFIG_SYS_MMC_MAX_DEVICE 2
+#define CONFIG_CMD_NET
+#define CONFIG_NET_MULTI

Signed-off-by: Che-Liang Chiou <clchiou@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agonet: dc2114x: check for apropriate command
Linus Walleij [Sun, 25 Sep 2011 21:41:14 +0000 (21:41 +0000)]
net: dc2114x: check for apropriate command

The code had two paths depending on whether the card was to be
accessed from plain memory or the IO region. However the error
path checks whether IO region was obtained - twice. Fix up the
error path according to the probable intention.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
13 years agokm/common: enable "env default" for all keymile boards
Gerlando Falauto [Thu, 22 Sep 2011 05:06:12 +0000 (05:06 +0000)]
km/common: enable "env default" for all keymile boards

Signed-off-by: Gerlando Falauto <gerlando.falauto@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
13 years agoftgmac100: reset fix when supports wake on lan
Macpaul Lin [Tue, 20 Sep 2011 19:54:32 +0000 (19:54 +0000)]
ftgmac100: reset fix when supports wake on lan

This patch move the reset function from initialization to
driver register procedure.

Some embedded system supports wake on lan nowadays. On this kind of system,
the ftgmac100 will be still supplied power after the system has been
shut-down by Linux. Hence the register used by linux won't be clear
when the system has been powered-off.

The origin ftgmac100 driver in u-boot will only register
driver and functions to network stack and won't reset the ftgmac100
hardware if the network won't be used during boot-up.
This will lead ftgmac100 continue receiving packets and then might corrupt
linux kernel when booting up.

So we reorder the hardware reset function earlier to force the hardware
to be reset whether it will be used or not.

Signed-off-by: Macpaul Lin <macpaul@andestech.com>
13 years agokm/common: remove obsolete header file
Holger Brunck [Tue, 13 Sep 2011 23:06:12 +0000 (23:06 +0000)]
km/common: remove obsolete header file

This header is unneeded due to code which was removed in the
past.

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Wolfgang Denk <wd@denx.de>

13 years agokm/common: fix ramfs development target
Andreas Huber [Tue, 13 Sep 2011 23:06:11 +0000 (23:06 +0000)]
km/common: fix ramfs development target

Calucations of PRAM needs to take into account the 'rootfssize'.
Memory available to the linux kernel 'mem=' is in all cases set to the total
memory size minus the pram size.

Signed-off-by: Andreas Huber <andreas.huber@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Wolfgang Denk <wd@denx.de>

13 years agokm/common: force set ethaddr only for KM kirkwood boards
Holger Brunck [Tue, 13 Sep 2011 23:06:10 +0000 (23:06 +0000)]
km/common: force set ethaddr only for KM kirkwood boards

This prevents u-boot to print out "Can't overwrite "ethaddr""
each time a powerpc board starts.

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Wolfgang Denk <wd@denx.de>

13 years agofix compile warning for env tools
Luka Perkov [Mon, 5 Sep 2011 21:40:13 +0000 (23:40 +0200)]
fix compile warning for env tools

Patch fixes this issue:

fw_env.c: In function ‘fw_setenv’:
fw_env.c:492:5: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 3 has type ‘size_t’ [-Wformat]
fw_env.c: In function ‘flash_write_buf’:
fw_env.c:806:6: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 3 has type ‘size_t’ [-Wformat]

Signed-off-by: Luka Perkov <lists@lukaperkov.net>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agokm_arm: enable POST for these boards
Valentin Longchamp [Fri, 2 Sep 2011 04:59:10 +0000 (04:59 +0000)]
km_arm: enable POST for these boards

The current km_arm boards have a Power-On test jumper. When this
jumper is set, this triggers some Power-On tests on the board.

This patch enables the support of this jumper for starting the
memory_regions test when the jumper is set.

Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Prafulla Wadaskar <prafulla@marvell.com>
13 years agoremoved static from images in cmd_bootm.c
Simon Schwarz [Fri, 2 Sep 2011 00:50:31 +0000 (00:50 +0000)]
removed static from images in cmd_bootm.c

This removes static modifier from images variable in cmd_bootm.c.

Signed-off-by: Simon Schwarz <simonschwarzcor@gmail.com>
13 years agoMerge branch 'master' of git://git.denx.de/u-boot-mpc85xx
Wolfgang Denk [Fri, 21 Oct 2011 21:48:46 +0000 (23:48 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-mpc85xx

* 'master' of git://git.denx.de/u-boot-mpc85xx:
  mpc85xx: Add inline GPIO acessor functions
  powerpc/85xx: wait for alignment before resetting SERDES RX lanes (SERDES9)
  powerpc/85xx: Fix P2020DS booting
  powerpc/85xx: Update USB device tree status based on pin settings
  fdt: Add new fdt_set_node_status & fdt_set_status_by_alias helpers
  powerpc/85xx: Add support for RMan LIODN initialization
  powerpc/85xx: Update device tree handling for SRIO
  powerpc/85xx: Update setting of SRIO LIODNs
  fm: Don't allow disabling of FM1-DTSEC1
  fm-eth: Don't mark the MAC we use for MDIO as disabled in device tree

13 years agoMerge branch 'master' of git://git.denx.de/u-boot-coldfire
Wolfgang Denk [Fri, 21 Oct 2011 21:36:40 +0000 (23:36 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-coldfire

* 'master' of git://git.denx.de/u-boot-coldfire:
  ColdFire: Clean Makefile _config rules
  ColdFire: Move boards with simple _config rules to boards.cfg
  ColdFire: Fix compilation with CONFIG_SYS_DRAMSZ1 defined
  ColdFire: Merge differentiated linking files into a sigle one by board
  ColdFire: Add $(obj) before cpu lib to correct build
  ColdFire: Cleanup lds files for multiple defined symbols

13 years agoMerge branch 'master' of /home/wd/git/u-boot/master
Wolfgang Denk [Fri, 21 Oct 2011 21:35:12 +0000 (23:35 +0200)]
Merge branch 'master' of /home/wd/git/u-boot/master

* 'master' of /home/wd/git/u-boot/master:
  cosmetic: Fixup fixup_silent_linux() for checkpatch
  Correct dependency rule to fix SPL build
  Move timestamp and version files into 'generated' subdir
  sandbox: Makefile changes to build sandbox architecture
  Add generic gpio.h in asm-generic
  Adjust dependency rules to permit per-file flags
  sandbox: Use uintptr_t for 32/64-bit compatibility
  sandbox: Add basic config file
  sandbox: Add serial uart
  sandbox: Add main program
  sandbox: Add OS dependent layer
  sandbox: Force command sections to be 4-byte aligned
  sandbox: Disable standalone/API support
  sandbox: Disable built-in malloc
  sandbox: Add bootm support
  sandbox: Add board info for architecture
  sandbox: Add sandbox board
  sandbox: Add architecture lib files
  sandbox: Add cpu files
  sandbox: Add compiler defines to support a 64-bit x86_64 platform
  sandbox: Add architecture image support
  Fix use of int as pointer in image.c
  sandbox: Add architecture header files
  arm: ca9x4_ct_vxp: enable PXE BOOTP options support
  arm: ca9x4_ct_vxp: enable pxe command support
  Convert ca9x4_ct_vxp to standard env variables
  net: bootp: add PXE/RFC 4578 DHCP options support
  Add pxe command
  lib: add uuid_str_to_bin for use with bootp and PXE uuid
  README: document standard image variables
  Replace space and tab checks with isblank
  cosmetic: remove unneeded curly braces
  Add isblank
  common: add run_command2 for running simple or hush commands
  common, menu: use abortboot for menu timeout
  Add generic, reusable menu code
  DM9000:Add a byte swap macro for dm9000 io operation.
  kw_gpio: fix error in kw_gpio_direction_input
  Blackfin: bfin_spi: fix build error when DEBUG is defined
  Blackfin: define CONFIG_SYS_CACHELINE_SIZE
  video: Moving mx3fb.c to CONFIG_VIDEO
  mx31: make HSP clock for mx3fb driver available
  MX5: Make IPU display output and pixel format configurable
  VIDEO: MX5: export pix format
  VIDEO: MX5: Switch MX5 to CONFIG_VIDEO
  video: update the Freescale DIU driver to use linux/fb.h
  powerpc: cpm2 boards: update fcc register logic

13 years agompc85xx: Add inline GPIO acessor functions
Kyle Moffett [Wed, 19 Oct 2011 18:58:15 +0000 (14:58 -0400)]
mpc85xx: Add inline GPIO acessor functions

To ease the implementation of other MPC85xx board ports, several common
GPIO helpers are added to <asm/mpc85xx_gpio.h>.

Since each of these compiles to no more than 4-5 instructions it would
be very inefficient to call them out of line, therefore we put them
entirely in the header file.

The HWW-1U-1A board port which these were written for strongly prefers
to set multiple GPIOs as a single batch operation, so the API is
designed around that basis.

To assist other board ports, a small set of wrappers are used which
provides a standard gpio_request() interface around the MPC85xx-specific
functions.  This can be enabled with CONFIG_MPC85XX_GENERIC_GPIO

Signed-off-by: Kyle Moffett <Kyle.D.Moffett@boeing.com>
Cc: Andy Fleming <afleming@gmail.com>
Cc: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agopowerpc/85xx: wait for alignment before resetting SERDES RX lanes (SERDES9)
Timur Tabi [Tue, 18 Oct 2011 23:44:34 +0000 (18:44 -0500)]
powerpc/85xx: wait for alignment before resetting SERDES RX lanes (SERDES9)

The work-around for P4080 erratum SERDES9 says that the SERDES receiver
lanes should be reset after the XAUI starts tranmitting alignment signals.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agocosmetic: Fixup fixup_silent_linux() for checkpatch
Doug Anderson [Wed, 19 Oct 2011 12:30:57 +0000 (12:30 +0000)]
cosmetic: Fixup fixup_silent_linux() for checkpatch

Signed-off-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Anton Staaf <robotboy@chromium.org>
13 years agoCorrect dependency rule to fix SPL build
Simon Glass [Tue, 18 Oct 2011 05:50:42 +0000 (05:50 +0000)]
Correct dependency rule to fix SPL build

Commit 4750884 introduced a change in the dependency generation which
breaks SPL, because the source files being built are not initially present
and are symlinked as part of the build.

The .depend file must depend not only on the files in the DEPS list but
also on the sources which did not contribute files to the DEPS list, since
these sources will otherwise not get a dependency and will not be built.

Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Wolfgang Denk <wd@denx.de>
13 years agoColdFire: Clean Makefile _config rules
stany MARCEL [Fri, 14 Oct 2011 04:38:05 +0000 (04:38 +0000)]
ColdFire: Clean Makefile _config rules

Remove unused mkdir, and add missing ones.

Signed-off-by: Stany MARCEL <stany.marcel@novasys-ingenierie.com>
13 years agoColdFire: Move boards with simple _config rules to boards.cfg
stany MARCEL [Tue, 18 Oct 2011 16:17:13 +0000 (00:17 +0800)]
ColdFire: Move boards with simple _config rules to boards.cfg

Signed-off-by: Stany MARCEL <stany.marcel@novasys-ingenierie.com>
13 years agoColdFire: Fix compilation with CONFIG_SYS_DRAMSZ1 defined
stany MARCEL [Fri, 14 Oct 2011 04:38:06 +0000 (04:38 +0000)]
ColdFire: Fix compilation with CONFIG_SYS_DRAMSZ1 defined

A temp variable was used but not declared, with CONFIG_SYS_DRAMSZ1
defined. This variable is now declared in the functione when needed.

Signed-off-by: Stany MARCEL <stany.marcel@novasys-ingenierie.com>
13 years agoColdFire: Merge differentiated linking files into a sigle one by board
stany MARCEL [Fri, 14 Oct 2011 04:38:04 +0000 (04:38 +0000)]
ColdFire: Merge differentiated linking files into a sigle one by board

The spa, stm, int, 32 and 16 linking files are identical so there is
no need to differentiate them. A single lds file is now used, and
_config rule are simplified.

Signed-off-by: Stany MARCEL <stany.marcel@novasys-ingenierie.com>
13 years agoColdFire: Add $(obj) before cpu lib to correct build
stany MARCEL [Fri, 14 Oct 2011 04:38:02 +0000 (04:38 +0000)]
ColdFire: Add $(obj) before cpu lib to correct build

Missing $(obj) prevented the build of ColdFire boards in a directory
than sources

Signed-off-by: Stany MARCEL <stany.marcel@novasys-ingenierie.com>
13 years agoColdFire: Cleanup lds files for multiple defined symbols
stany MARCEL [Fri, 14 Oct 2011 04:38:01 +0000 (04:38 +0000)]
ColdFire: Cleanup lds files for multiple defined symbols

Lds files cleened to remove multiple defined section and modified to
be compliant with --gc-sections added for ColdFire platform in a
previous patch.

Signed-off-by: Stany MARCEL <stany.marcel@novasys-ingenierie.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agopowerpc/85xx: Fix P2020DS booting
Kumar Gala [Tue, 18 Oct 2011 06:31:08 +0000 (01:31 -0500)]
powerpc/85xx: Fix P2020DS booting

The following commit removed the code that set odt_rd_cfg and
odt_wr_cfg.  With out this code P2020DS board will not boot:

commit 712cf7ab0b58e51a69e339397457d3591b6b650e
Author: York Sun <yorksun@freescale.com>
Date:   Mon Oct 3 09:19:53 2011 -0700

    powerpc/mpc8xxx: Merge entries in DDR speed table

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agopowerpc/85xx: Update USB device tree status based on pin settings
Shengzhou Liu [Fri, 14 Oct 2011 08:26:06 +0000 (16:26 +0800)]
powerpc/85xx: Update USB device tree status based on pin settings

For P3060 and P4080, USB pins are multiplexed with other functions.
Update the device tree status for USB ports based on setting of
RCW[EC1] & RCW[EC2] which describe if pins are muxed to usb.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agofdt: Add new fdt_set_node_status & fdt_set_status_by_alias helpers
Shengzhou Liu [Fri, 14 Oct 2011 08:26:05 +0000 (16:26 +0800)]
fdt: Add new fdt_set_node_status & fdt_set_status_by_alias helpers

Add common function fdt_set_node_status() to assist in various locations
that we set a nodes status.  This function utilizes the status values
that are part of the EPAPR spec (on power.org).

fdt_set_status_by_alias() is based on fdt_set_node_status() but uses an
alias string to identify the node to update.

We also add some shortcut functions to help the common cases of setting
"okay" and "disabled":

fdt_status_okay()
fdt_status_disabled()
fdt_status_okay_by_alias()
fdt_status_disabled_by_alias()

Finally, we fixup the corenet_ds ethernet code which previously had
a function by the same name that can be replaced with the new helpers.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Gerald Van Baren <vanbaren@cideas.com>
13 years agopowerpc/85xx: Add support for RMan LIODN initialization
Kumar Gala [Fri, 14 Oct 2011 18:28:52 +0000 (13:28 -0500)]
powerpc/85xx: Add support for RMan LIODN initialization

This patch is intended to initialize RMan LIODN related registers on
P2041, P304S and P5020 SocS. It also adds the "rman@0" child node to
qman-portal nodes, adds "fsl,liodn" property to RMan inbound block nodes.

Signed-off-by: Minghuan Lian <Minghuan.Lian@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agopowerpc/85xx: Update device tree handling for SRIO
Kumar Gala [Fri, 14 Oct 2011 05:03:58 +0000 (00:03 -0500)]
powerpc/85xx: Update device tree handling for SRIO

Update device tree handling for SRIO controller to support updated
fsl,srio device tree binding.

We handle disabling of individual ports, the whole controller, RMU, and
RMAN.  Additionally, we setup the SRIO related LIODNs in the device
tree.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agopowerpc/85xx: Update setting of SRIO LIODNs
Kumar Gala [Fri, 14 Oct 2011 05:01:23 +0000 (00:01 -0500)]
powerpc/85xx: Update setting of SRIO LIODNs

Properly set the LIODN values associated with SRIO controller.  On
P4080/P3060 we have an LIODN per port and one for the RMU.  On
P2041/P3041/P5020 we have 2 LIODNs per port.

Update the tables for all of these devices to properly handle both
styles.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agofm: Don't allow disabling of FM1-DTSEC1
Kumar Gala [Fri, 14 Oct 2011 08:17:56 +0000 (03:17 -0500)]
fm: Don't allow disabling of FM1-DTSEC1

The MDIO controller to talk to external PHYs is on FM1-DTSEC1 so don't
allow disabling.  If we disable it we end up powering the block down in
the SoC and thus can't communicate to any external PHYs.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agofm-eth: Don't mark the MAC we use for MDIO as disabled in device tree
Kumar Gala [Thu, 13 Oct 2011 19:55:59 +0000 (14:55 -0500)]
fm-eth: Don't mark the MAC we use for MDIO as disabled in device tree

FM1-DTSEC1's MAC was being marked as disabled if the port was not
configured based on the SoC configuration.  However we utilize the MAC
interface for MDIO and thus should NOT mark it disabled.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agoMove timestamp and version files into 'generated' subdir
Simon Glass [Mon, 10 Oct 2011 08:55:19 +0000 (08:55 +0000)]
Move timestamp and version files into 'generated' subdir

There is a rather subtle build problem where the build time stamp is not
updated for out-of-tree builds if there exists an in-tree build which
has a valid timestamp file. So if you do an in-tree build, then an
out-of-tree build your timestamp will not change.

The correct timestamp_autogenerated.h lives in the object tree, but it
is not always found there. The source still lives in the source tree and
when compiling version.h, it includes timestamp_autogenerated.h. Since
the current directory is always searched first, this will come from the
source tree rather than the object tree if it exists there. This affects
dependency generation also, which means that common/cmd_version.o will not
even be rebuilt if you have ever done an in-tree build.

A similar problem exists with the version file.

This change moves both files into the 'generated' subdir, which is already
used for asm-offsets.h. Then timestamp.h and version.h are updated to
include the files from there.

There are other places where these generated files are included, but I
cannot see why these don't just use the timestamp.h and version.h headers.
So this change also tidies that up.

I have tested this with in- and out-of-tree builds, but not SPL. I have
looked at various other options for fixing this, including sed on the dep
files, -I- and -include flags to gcc, but I don't think they can be made
to work. Comments welcome.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agosandbox: Makefile changes to build sandbox architecture
Simon Glass [Mon, 3 Oct 2011 19:26:50 +0000 (19:26 +0000)]
sandbox: Makefile changes to build sandbox architecture

At this point U-Boot will build and run on x86 under Linux.

The idea is to define a new architecture called 'sandbox', alongside ARM
and x86. This runs natively on Linux to suit the host machine. All
hardware access is either omitted or emulated.

The purpose of this system is to test the bulk of the non-hardware-specific
U-Boot code. We can mock the SPI flash, GPIOs, UART and keyboard, then test
that U-Boot behaves as we wish.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agoAdd generic gpio.h in asm-generic
Simon Glass [Fri, 7 Oct 2011 13:53:50 +0000 (13:53 +0000)]
Add generic gpio.h in asm-generic

Since we want want to have a standard GPIO interface, this adds a definition
for this into include/asm-generic/gpio.h.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agoAdjust dependency rules to permit per-file flags
Simon Glass [Fri, 7 Oct 2011 13:53:49 +0000 (13:53 +0000)]
Adjust dependency rules to permit per-file flags

The dependency rules are currently done in a shell 'for' loop. This does not
permit Makefile variables to adjust preprocessor flags as is done with normal
compile flags, using the CFLAGS_path/file.o syntax.

This change moves the dependency generation into the Makefile itself, and
permits a CPPFLAGS_path/file.o to adjust preprocessor flags on a file or
directory basis.

The CPPFLAGS_... variable is also folded into CFLAGS during the build.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Use uintptr_t for 32/64-bit compatibility
Simon Glass [Sat, 17 Sep 2011 06:48:58 +0000 (06:48 +0000)]
sandbox: Use uintptr_t for 32/64-bit compatibility

This fixes a problems when building on some 64-bit machines.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add basic config file
Simon Glass [Mon, 3 Oct 2011 19:26:47 +0000 (19:26 +0000)]
sandbox: Add basic config file

This basic provides required features along with a basic command set.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add serial uart
Simon Glass [Mon, 3 Oct 2011 19:26:46 +0000 (19:26 +0000)]
sandbox: Add serial uart

This uart simply writes to stdout and reads from stdin. We might imagine
instead buffering the data so that a test interface can check output and
inject input.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add main program
Simon Glass [Mon, 3 Oct 2011 19:26:45 +0000 (19:26 +0000)]
sandbox: Add main program

Add a main program so that we can run U-Boot.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add OS dependent layer
Simon Glass [Mon, 3 Oct 2011 19:26:44 +0000 (19:26 +0000)]
sandbox: Add OS dependent layer

We want to keep all OS-dependent code in once place, with a simple interface
to U-Boot. For now, this is that place.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Force command sections to be 4-byte aligned
Simon Glass [Sat, 17 Sep 2011 06:48:51 +0000 (06:48 +0000)]
sandbox: Force command sections to be 4-byte aligned

By default sections are 16-byte aligned on some architectures, but the
command name structure (struct cmd_tbl_s) does not have padding to
16 bytes. This reduces the alignment to 4-bytes so that the command
table can be accessed correctly on any architecture.

(Note: this needs doing properly)

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Disable standalone/API support
Simon Glass [Sat, 17 Sep 2011 06:48:50 +0000 (06:48 +0000)]
sandbox: Disable standalone/API support

This is not useful on the sandbox architecture since we can simply link all
our code with U-Boot.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Disable built-in malloc
Simon Glass [Sat, 17 Sep 2011 06:48:49 +0000 (06:48 +0000)]
sandbox: Disable built-in malloc

We prefer to U-Boot's malloc but for now it is easier to use the C library's
version.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add bootm support
Simon Glass [Fri, 7 Oct 2011 13:53:40 +0000 (13:53 +0000)]
sandbox: Add bootm support

This adds sandbox architecture support to bootm, although it is probably
not useful to load sandbox code into the address space and execute it.

This change at least make the file build correctly on 64-bit machines.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add board info for architecture
Simon Glass [Sat, 17 Sep 2011 06:48:47 +0000 (06:48 +0000)]
sandbox: Add board info for architecture

This is required for the bdinfo command to work.

Signed-off-by: Simon Glass <sjg@chromium.org>
Fix syntax error.
Signed-off-by: Wolfgang Denk <wd@denx.de>
13 years agosandbox: Add sandbox board
Simon Glass [Fri, 7 Oct 2011 13:53:38 +0000 (13:53 +0000)]
sandbox: Add sandbox board

This adds basic files for the sandbox board. The lds file is very simple
since we can rely mostly on the linker defaults.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add architecture lib files
Simon Glass [Mon, 3 Oct 2011 19:26:37 +0000 (19:26 +0000)]
sandbox: Add architecture lib files

These files are taken from the ARM board implementation and then reduced
to remove unneeded cruft.

Ideally we would work towards unifying arch/xxx/lib files, particularly
board.c.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add cpu files
Simon Glass [Mon, 26 Sep 2011 14:10:39 +0000 (14:10 +0000)]
sandbox: Add cpu files

This is an initial implementation with all functions defined but not working.

The lds file is very simple since we can mostly rely on the linker defaults.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add compiler defines to support a 64-bit x86_64 platform
Simon Glass [Fri, 23 Sep 2011 06:22:06 +0000 (06:22 +0000)]
sandbox: Add compiler defines to support a 64-bit x86_64 platform

This sets __WORDSIZE to 8 correctly on 64-bit machines.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agosandbox: Add architecture image support
Simon Glass [Fri, 7 Oct 2011 13:53:34 +0000 (13:53 +0000)]
sandbox: Add architecture image support

We won't actually load an image with this architecture, but we still need to
define it.

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agoFix use of int as pointer in image.c
Simon Glass [Fri, 23 Sep 2011 06:22:04 +0000 (06:22 +0000)]
Fix use of int as pointer in image.c

It is better to use %p in this case.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agosandbox: Add architecture header files
Simon Glass [Mon, 10 Oct 2011 08:22:14 +0000 (08:22 +0000)]
sandbox: Add architecture header files

This adds required header files for the sandbox architecture, and a basic
description of what sandbox is (README.sandbox).

Signed-off-by: Simon Glass <sjg@chromium.org>
13 years agoarm: ca9x4_ct_vxp: enable PXE BOOTP options support
Jason Hobbs [Tue, 23 Aug 2011 11:07:01 +0000 (11:07 +0000)]
arm: ca9x4_ct_vxp: enable PXE BOOTP options support

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
Cc: Matt Waddel <matt.waddel@linaro.org>
13 years agoarm: ca9x4_ct_vxp: enable pxe command support
Jason Hobbs [Tue, 23 Aug 2011 11:07:00 +0000 (11:07 +0000)]
arm: ca9x4_ct_vxp: enable pxe command support

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
Cc: Matt Waddel <matt.waddel@linaro.org>
13 years agoConvert ca9x4_ct_vxp to standard env variables
Jason Hobbs [Tue, 23 Aug 2011 11:06:59 +0000 (11:06 +0000)]
Convert ca9x4_ct_vxp to standard env variables

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
Cc: Matt Waddel <matt.waddel@linaro.org>
13 years agonet: bootp: add PXE/RFC 4578 DHCP options support
Jason Hobbs [Wed, 31 Aug 2011 05:37:31 +0000 (05:37 +0000)]
net: bootp: add PXE/RFC 4578 DHCP options support

These options are required to be present in RFC 4578 compliant DHCP
requests. They give more information to DHCP servers to allow serving

different DHCP responses to different systems based on client
architecture, client capabilities, UUID, or vendor.

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agoAdd pxe command
Jason Hobbs [Wed, 31 Aug 2011 15:37:30 +0000 (10:37 -0500)]
Add pxe command

Add pxe command, which is intended to mimic PXELINUX functionality.
'pxe get' uses tftp to retrieve a file based on UUID, MAC address or IP
address. 'pxe boot' interprets the contents of PXELINUX config like file
to boot using a specific initrd, kernel and kernel command line.

This patch also adds a README.pxe file - see it for more details on the
pxe command.

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agolib: add uuid_str_to_bin for use with bootp and PXE uuid
Jason Hobbs [Tue, 23 Aug 2011 11:06:56 +0000 (11:06 +0000)]
lib: add uuid_str_to_bin for use with bootp and PXE uuid

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agoREADME: document standard image variables
Jason Hobbs [Wed, 31 Aug 2011 05:37:28 +0000 (05:37 +0000)]
README: document standard image variables

With these documented, we can start pushing towards standardizing their
use across boards.

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
Cc: Mike Frysinger <vapier@gentoo.org>
13 years agoReplace space and tab checks with isblank
Jason Hobbs [Tue, 23 Aug 2011 11:06:54 +0000 (11:06 +0000)]
Replace space and tab checks with isblank

These are various places I found that checked for conditions equivalent
to isblank.

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agocosmetic: remove unneeded curly braces
Jason Hobbs [Tue, 23 Aug 2011 11:06:53 +0000 (11:06 +0000)]
cosmetic: remove unneeded curly braces

This prevents a checkpatch warning in the patch to use isblank

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agoAdd isblank
Jason Hobbs [Tue, 23 Aug 2011 11:06:52 +0000 (11:06 +0000)]
Add isblank

Existing ctype checks are implemented using a 256 byte lookup table,
allowing each character to be in any of 8 character classes. Since there
are 8 existing character classes without the blank class, I implemented
isblank without using the lookup table.  Since there are only two blank
characters - tab and space - this is a more reasonable approach than
doubling the size of the lookup table to accommodate one more class.

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agocommon: add run_command2 for running simple or hush commands
Jason Hobbs [Wed, 31 Aug 2011 05:37:24 +0000 (05:37 +0000)]
common: add run_command2 for running simple or hush commands

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
Cc: Mike Frysinger <vapier@gentoo.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agocommon, menu: use abortboot for menu timeout
Jason Hobbs [Tue, 23 Aug 2011 11:06:50 +0000 (11:06 +0000)]
common, menu: use abortboot for menu timeout

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agoAdd generic, reusable menu code
Jason Hobbs [Tue, 23 Aug 2011 11:06:49 +0000 (11:06 +0000)]
Add generic, reusable menu code

This will be used first by the pxe code, but is intended to be
generic and reusable for other jobs in U-boot.

Signed-off-by: Jason Hobbs <jason.hobbs@calxeda.com>
13 years agoDM9000:Add a byte swap macro for dm9000 io operation.
Jason Jin [Thu, 25 Aug 2011 07:46:43 +0000 (15:46 +0800)]
DM9000:Add a byte swap macro for dm9000 io operation.

commit a45dde2293c816138e53c26eca6fd0322583f9a6 changed the dm9000
direct register access to standard IO. This should work
on the ColdFire platform as there are corresponding macros for
the LE devices. But the hardware settings on some ColdFire boards had
swapped the byte order which make the original macros such as out_le16
cannot work. To avoid changing the common io access code on ColdFire
platform, the DM9000_BYTE_SWAPPED define was added to make the dm9000 use
__raw* IO access on some ColdFire boards.

Signed-off-by: Jason Jin <Jason.jin@freescale.com>
13 years agokw_gpio: fix error in kw_gpio_direction_input
Holger Brunck [Sun, 21 Aug 2011 23:21:06 +0000 (23:21 +0000)]
kw_gpio: fix error in kw_gpio_direction_input

The function kw_gpio_is_valid returns zero on success, so
adapt the error check accordingly.

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
cc: Dieter Kiermaier <dk-arm-linux@gmx.de>

13 years agoBlackfin: bfin_spi: fix build error when DEBUG is defined
Mike Frysinger [Tue, 4 Oct 2011 00:51:13 +0000 (20:51 -0400)]
Blackfin: bfin_spi: fix build error when DEBUG is defined

The new debug() changes pointed out debug code that rotted a little.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
13 years agoBlackfin: define CONFIG_SYS_CACHELINE_SIZE
Mike Frysinger [Thu, 13 Oct 2011 00:16:31 +0000 (20:16 -0400)]
Blackfin: define CONFIG_SYS_CACHELINE_SIZE

Common U-Boot API wants this define, so import asm/cache.h from Linux
to provide suitable defines.

Acked-by: Anton Staaf <robotboy@chromium.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
13 years agovideo: Moving mx3fb.c to CONFIG_VIDEO
Helmut Raiger [Wed, 12 Oct 2011 23:16:29 +0000 (23:16 +0000)]
video: Moving mx3fb.c to CONFIG_VIDEO

mx3fb.c was based on CONFIG_LCD and is moved by this patch to
CONFIG_VIDEO, which has greater freedom in selecting videomodes
even at runtime.

This renders the accumulating list of display defines
(CONFIG_DISPLAY_VBEST..., CONFIG_DISPLAY_C057...) obsolete as
these may be setup through env variables:

uboot> setenv mydisplay 'video=ctfb:x:240,y:320,depth:16,mode:0,pclk:185925,
le:9,ri:17,up:7,lo:10,hs:1,vs:1,sync:100663296,vmode:0'
uboot> setenv videomode ${mydisplay}

This commit also fixes the board config files for qong and
imx31_phycore boards as needed. The videomode settings of
previously supported displays are added to CONFIG_EXTRA_ENV_SETTINGS
now. CONFIG_SYS_MALLOC_LEN for imx31_phycore board is increased
to make the frame buffer allocation working with the changed
driver.

Signed-off-by: Helmut Raiger <helmut.raiger@hale.at>
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
13 years agomx31: make HSP clock for mx3fb driver available
Helmut Raiger [Wed, 12 Oct 2011 21:08:30 +0000 (23:08 +0200)]
mx31: make HSP clock for mx3fb driver available

This additionally updates mx31/generic.c by
- replacing __REG() macro accesses with readl() and writel()
- providing macros for PDR0 and PLL bit accesses

Signed-off-by: Helmut Raiger <helmut.raiger@hale.at>
Acked-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Stefano Babic <sbabic@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
Signed-off-by: Anatolij Gustschin <agust@denx.de>
13 years agoMX5: Make IPU display output and pixel format configurable
Marek Vasut [Wed, 5 Oct 2011 22:25:03 +0000 (00:25 +0200)]
MX5: Make IPU display output and pixel format configurable

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Stefano Babic <sbabic@denx.de>
Tested-by: Stefano Babic <sbabic@denx.de>
Signed-off-by: Anatolij Gustschin <agust@denx.de>
13 years agoVIDEO: MX5: export pix format
Stefano Babic [Wed, 28 Sep 2011 09:21:16 +0000 (11:21 +0200)]
VIDEO: MX5: export pix format

Pixel format defines must be available for boards to set up
the right display. Move them and export in a new file.

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Anatolij Gustschin <agust@denx.de>
CC: Marek Vasut <marek.vasut@gmail.com>
Tested-by: Stefano Babic <sbabic@denx.de>
Signed-off-by: Anatolij Gustschin <agust@denx.de>
13 years agoVIDEO: MX5: Switch MX5 to CONFIG_VIDEO
Stefano Babic [Wed, 28 Sep 2011 09:21:15 +0000 (11:21 +0200)]
VIDEO: MX5: Switch MX5 to CONFIG_VIDEO

The framebuffer driver for MX5 is based on CONFIG_LCD.
In the current implementation, there is a serious bug
because the required memory is allocated before
relocation, but the driver knows only later which is
the resolution of the display. The patch switches the driver
to CONFIG_VIDEO and the memory is allocated by the driver itself.

We also need to switch the vision2 board code and config file
in the same commit so that this commit will be bisectable.

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Anatolij Gustschin <agust@denx.de>
Tested-by: Stefano Babic <sbabic@denx.de>
Signed-off-by: Anatolij Gustschin <agust@denx.de>
13 years agovideo: update the Freescale DIU driver to use linux/fb.h
Timur Tabi [Mon, 22 Aug 2011 15:54:02 +0000 (10:54 -0500)]
video: update the Freescale DIU driver to use linux/fb.h

Update the Freescale DIU video driver (fsl_diu_fb.c) to use linux/fb.h.
Some data structures from this header file were just copied into
fsl_diu_fb.c.

Signed-off-by: Timur Tabi <timur@freescale.com>
13 years agopowerpc: cpm2 boards: update fcc register logic
Mike Frysinger [Mon, 17 Oct 2011 05:38:58 +0000 (05:38 +0000)]
powerpc: cpm2 boards: update fcc register logic

In the recent dropping of !NET_MULTI code (commit e2a53458a7ab37523304),
I misread the logic in include/net.h.  Some of it was used by NET_MULTI
code as glue between the multi/non-multi worlds for cpm2 boards.

Rather than restore the block of code, push the logic to the board config
headers where it all belongs.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
13 years agoMerge branch 'master' of git://git.denx.de/u-boot-blackfin
Wolfgang Denk [Mon, 17 Oct 2011 19:50:52 +0000 (21:50 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-blackfin

* 'master' of git://git.denx.de/u-boot-blackfin:
  Blackfin: bfin_spi: fix build error when DEBUG is defined
  Blackfin: define CONFIG_SYS_CACHELINE_SIZE

13 years agoMerge branch 'master' of git://git.denx.de/u-boot-video
Wolfgang Denk [Mon, 17 Oct 2011 19:48:20 +0000 (21:48 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-video

* 'master' of git://git.denx.de/u-boot-video:
  video: Moving mx3fb.c to CONFIG_VIDEO
  mx31: make HSP clock for mx3fb driver available
  MX5: Make IPU display output and pixel format configurable
  VIDEO: MX5: export pix format
  VIDEO: MX5: Switch MX5 to CONFIG_VIDEO
  video: update the Freescale DIU driver to use linux/fb.h

13 years agozlib: handle overflow while calculating available stream input size
Anatolij Gustschin [Sat, 15 Oct 2011 06:32:52 +0000 (06:32 +0000)]
zlib: handle overflow while calculating available stream input size

If compressed data is located in sectors at the end of the flash and
it's offset + input stream size > 0xFFFFFFFF, the uncompressing time
is very long, since processing of the stream is done bytewise (and
not blockwise) due to overflow in inflate_fast() while calculation
and checking for enough input available.

Check for this overflow condition and limit the available stream
input size to the actually max. possible input size. This fixes
the problem.

The issue is easily reproduceable by placing a gziped bitmap in flash,
e.g. at FFF80000, and running 'bmp' commands like 'bmp info FFF80000'
or 'bmp display FFF80000'. The uncompressing can take up to 3 sec.
whereas it should normaly take a fraction of a second. If the
'splashimage' environment variable points to this address, the
booting time also increases significantly.

Signed-off-by: Anatolij Gustschin <agust@denx.de>
13 years agoBlackfin: bfin_spi: fix build error when DEBUG is defined
Mike Frysinger [Tue, 4 Oct 2011 00:51:13 +0000 (20:51 -0400)]
Blackfin: bfin_spi: fix build error when DEBUG is defined

The new debug() changes pointed out debug code that rotted a little.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
13 years agoBlackfin: define CONFIG_SYS_CACHELINE_SIZE
Mike Frysinger [Thu, 13 Oct 2011 00:16:31 +0000 (20:16 -0400)]
Blackfin: define CONFIG_SYS_CACHELINE_SIZE

Common U-Boot API wants this define, so import asm/cache.h from Linux
to provide suitable defines.

Acked-by: Anton Staaf <robotboy@chromium.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
13 years agoserial_exit: punt unused prototype
Mike Frysinger [Mon, 3 Oct 2011 15:11:14 +0000 (15:11 +0000)]
serial_exit: punt unused prototype

No code defines or calls this, so drop the prototype.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Simon Glass <sjg@chromium.org>
13 years agopunt unused clean/distclean targets
Mike Frysinger [Thu, 13 Oct 2011 06:54:57 +0000 (06:54 +0000)]
punt unused clean/distclean targets

The top level Makefile does not do any recursion into subdirs when
cleaning, so these clean/distclean targets in random arch/board dirs
never get used.  Punt them all.

MAKEALL didn't report any errors related to this that I could see.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
13 years agocommon: fix missing function pointer relocation in fixup_cmdtable()
Daniel Schwierzeck [Thu, 13 Oct 2011 06:07:39 +0000 (06:07 +0000)]
common: fix missing function pointer relocation in fixup_cmdtable()

The command auto-completion does not work on architectures relying
on CONFIG_NEEDS_MANUAL_RELOC like MIPS. Cause is the missing function
pointer fixup for cmd_tbl_t::complete function in fixup_cmdtable().

This patch adds the missing pointer fixup in case of CONFIG_AUTO_COMPLETE
is defined.

Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
13 years agompc5200: digsy_mtc: fix detection of Coral-PA
Anatolij Gustschin [Thu, 13 Oct 2011 05:19:17 +0000 (05:19 +0000)]
mpc5200: digsy_mtc: fix detection of Coral-PA

A delay of approximately 250 ms after PCI bus reset in
pci_mpc5xxx_init() is needed to recognize the Coral-PA
controller on the graphic extention board.

Signed-off-by: Anatolij Gustschin <agust@denx.de>
Acked-by: Detlev Zundel <dzu@denx.de>
13 years agopci: move pcidelay code to new location just before PCI bus scan
Anatolij Gustschin [Tue, 11 Oct 2011 22:44:30 +0000 (22:44 +0000)]
pci: move pcidelay code to new location just before PCI bus scan

PCI cards might need some time after reset to respond. On some
boards (mpc5200 or mpc8260 based) the PCI bus reset is deasserted
at pci_init_board() time, so we currently can not use available
"pcidelay" option for waiting before PCI bus scan since this
waiting takes place before calling pci_init_board(). By moving
the pcidelay code to the new location using of the "pcidelay"
option is possible on mpc5200 or mpc8260 based boards, too.

Since pci_hose_scan() could be called multiple times, restrict
the function to wait only during its first call and to ignore
pcidelay for any further call (as pointed out by Matthias).

Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: Matthias Fuchs <matthias.fuchs@esd.eu>
Acked-by: Stefan Roese <sr@denx.de>
Acked-by: Matthias Fuchs <matthias.fuchs@esd.eu>
Tested-by: Matthias Fuchs <matthias.fuchs@esd.eu>
13 years agonet: axi_ethernet: Add driver to u-boot
Michal Simek [Thu, 6 Oct 2011 20:35:35 +0000 (20:35 +0000)]
net: axi_ethernet: Add driver to u-boot

Add axi_ethernet driver for little-endian Microblaze.

RX/TX BDs and rxframe buffer are shared among all axi_ethernet MACs.
Only one MAC can work in one time.

Signed-off-by: Michal Simek <monstr@monstr.eu>
Acked-by: Mike Frysinger <vapier@gentoo.org>
13 years agonet: emaclite: Fix coding style
Michal Simek [Mon, 12 Sep 2011 21:10:05 +0000 (21:10 +0000)]
net: emaclite: Fix coding style

Coding style should follow linux coding style.

Signed-off-by: Michal Simek <monstr@monstr.eu>
13 years agonet: emaclite: Use PKTSIZE directly
Michal Simek [Mon, 12 Sep 2011 21:10:04 +0000 (21:10 +0000)]
net: emaclite: Use PKTSIZE directly

Do not setup additional ENET_MAX_MTU macro.

Signed-off-by: Michal Simek <monstr@monstr.eu>
13 years agonet: emaclite: Setup RX/TX ping pong for every instance
Michal Simek [Mon, 12 Sep 2011 21:10:01 +0000 (21:10 +0000)]
net: emaclite: Setup RX/TX ping pong for every instance

Setup RX/TX ping-pong buffer for every emaclite IP separately.
The next patch move initialization directly to board code.

Signed-off-by: Michal Simek <monstr@monstr.eu>
13 years agoMerge branch 'master' of git://git.denx.de/u-boot-fdt
Wolfgang Denk [Sat, 15 Oct 2011 20:00:01 +0000 (22:00 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-fdt

* 'master' of git://git.denx.de/u-boot-fdt:
  powerpc/85xx: use fdt_create_phandle() to create the Fman firmware phandles
  fdt: update fdt_alloc_phandle to use fdt_get_phandle
  fdt: check for fdt errors in fdt_create_phandle
  fdt: Add a do_fixup_by_path_string() function

13 years agopowerpc/85xx: use fdt_create_phandle() to create the Fman firmware phandles
Timur Tabi [Tue, 20 Sep 2011 23:24:36 +0000 (18:24 -0500)]
powerpc/85xx: use fdt_create_phandle() to create the Fman firmware phandles

Function fdt_create_phandle() conveniently creates new phandle properties
using both "linux,phandle" and "phandle", so it should be used by all code
that wants to create a phandle.

The Fman firmware code, which embeds an Fman firmware into the device tree,
was creating the phandle properties manually.  Instead, change it to use
fdt_create_phandle().

Signed-off-by: Timur Tabi <timur@freescale.com>
13 years agofdt: update fdt_alloc_phandle to use fdt_get_phandle
Timur Tabi [Tue, 20 Sep 2011 23:24:35 +0000 (18:24 -0500)]
fdt: update fdt_alloc_phandle to use fdt_get_phandle

The device tree compiler, dtc, can use "phandle" and/or "linux,phandle"
properties to specify the phandle for any node.  By default, it uses
both, but "linux,phandle" is deprecated.  One day, we'd like to stop using
"linux,phandle", but U-boot needs to support both properties equally
first.

fdt_alloc_phandle() generates a unique phandle, but it was only checking
the "linux,phandle" properties.  Instead, we use fdt_get_phandle(),
which checks both properties automatically.  This ensures that we
support dtbs that only use "phandle".

The side-effect is that fdt_alloc_phandle() now takes twice as long, since
it has to check for two properties instead of one in each node that it
searches.

Signed-off-by: Timur Tabi <timur@freescale.com>
13 years agofdt: check for fdt errors in fdt_create_phandle
Timur Tabi [Tue, 20 Sep 2011 23:24:34 +0000 (18:24 -0500)]
fdt: check for fdt errors in fdt_create_phandle

fdt_create_phandle() was ignoring errors from fdt_set_phandle().  If an
error occurs, print an error message and return 0, which is an invalid
phandle.  We also need to change the return type for fdt_create_phandle()
to indicate that it cannot return an error code.

Signed-off-by: Timur Tabi <timur@freescale.com>
13 years agofdt: Add a do_fixup_by_path_string() function
Chunhe Lan [Wed, 31 Aug 2011 02:36:32 +0000 (10:36 +0800)]
fdt: Add a do_fixup_by_path_string() function

The do_fixup_by_path_string() will set the specified node's property to the
value contained in "status". It would just be an inline wrapper for
do_fixup_by_path() that calls strlen on the argument.

Signed-off-by: Chunhe Lan <Chunhe.Lan@freescale.com>
13 years agopowerpc/85xx: Drop CONFIG_VIDEO on P1_P2_RDB-PC boards to reduce size
Kumar Gala [Wed, 12 Oct 2011 20:52:08 +0000 (15:52 -0500)]
powerpc/85xx: Drop CONFIG_VIDEO on P1_P2_RDB-PC boards to reduce size

With older compilers (gcc-4.2.x) we run into issues that resulting image
is too large.  We can save a bunch of space by removing the video support.
In general video support on these boards is a nice to have since it
requires a PCIe add-on card.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
13 years agophylib: wait for TN2020 to achieve SERDES lane alignment at startup
Timur Tabi [Fri, 7 Oct 2011 20:35:21 +0000 (15:35 -0500)]
phylib: wait for TN2020 to achieve SERDES lane alignment at startup

Before the Teranetics TN2020 PHY can be used, the SERDES lanes need to be
aligned, so wait for lane alignment before completing the startup sequence.

Note that this process can take up to three seconds.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>