oweals/u-boot.git
9 years agoomap: SPL boot devices cleanup and completion
Paul Kocialkowski [Wed, 15 Jul 2015 14:02:20 +0000 (16:02 +0200)]
omap: SPL boot devices cleanup and completion

This cleans up the SPL boot devices for omap platforms and introduces support
for missing boot devices.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agoomap-common: Common boot code OMAP3 support and cleanup
Paul Kocialkowski [Wed, 15 Jul 2015 14:02:19 +0000 (16:02 +0200)]
omap-common: Common boot code OMAP3 support and cleanup

This introduces OMAP3 support for the common omap boot code, as well as a
major cleanup of the common omap boot code.

First, the omap_boot_parameters structure becomes platform-specific, since its
definition differs a bit across omap platforms. The offsets are removed as well
since it is U-Boot's coding style to use structures for mapping such kind of
data (in the sense that it is similar to registers). It is correct to assume
that romcode structure encoding is the same as U-Boot, given the description
of these structures in the TRMs.

The original address provided by the bootrom is passed to the U-Boot binary
instead of a duplicate of the structure stored in global data. This allows to
have only the relevant (boot device and mode) information stored in global data.
It is also expected that the address where the bootrom stores that information
is not overridden by the U-Boot SPL or U-Boot.

The save_omap_boot_params is expected to handle all special cases where the data
provided by the bootrom cannot be used as-is, so that spl_boot_device and
spl_boot_mode only return the data from global data.

All of this is only relevant when the U-Boot SPL is used. In cases it is not,
save_boot_params should fallback to its weak (or board-specific) definition.
save_omap_boot_params should not be called in that context either.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agokbuild: create symbolic link only for ARM, AVR32, SPARC, PowerPC, x86
Masahiro Yamada [Wed, 15 Jul 2015 11:59:29 +0000 (20:59 +0900)]
kbuild: create symbolic link only for ARM, AVR32, SPARC, PowerPC, x86

The symbolic link to SoC/CPU specific header directory is created
during the build, while it is only necessary for ARM, AVR32, SPARC,
x86, and some CPUs of PowerPC.  For the other architectures, it just
results in a broken symbolic link.

Introduce CONFIG_CREATE_ARCH_SYMLINK to not create unneeded symbolic
links.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
9 years agonds32: include <asm/arch-*/*.h> instead of <asm/arch/*.h>
Masahiro Yamada [Wed, 15 Jul 2015 11:59:28 +0000 (20:59 +0900)]
nds32: include <asm/arch-*/*.h> instead of <asm/arch/*.h>

There are only two SoC-specific headers for this architecture:
 - arch/nds32/include/asm/arch-ag101/ag101.h
 - arch/nds32/include/asm/arch-ag102/ag102.h

Those two have different file names, so there is no advantage to
include them via symbolic linked directory.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agogeneric_board: Call "checkboard" even though the root node has a "model" property
Haikun.Wang@freescale.com [Thu, 9 Jul 2015 11:58:03 +0000 (19:58 +0800)]
generic_board: Call "checkboard" even though the root node has a "model" property

In case of enable CONFIG_OF_CONTROL and has a "model" property in the root node,
the board special "checkboard" will not be called.
Usually we show some useful version information in the function.
This patch enable call "checkboard" in any case.
It is not conflicting with showing "model" at the same time.

For example on LS2085AQDS:
Showing "model" only:
Model: Freescale Layerscape 2085a QDS Board

Showing "checkboard" only:
Board: LS2085E-QDS, Board Arch: V1, Board version: B, boot from vBank: 4

Showing both:
Model: Freescale Layerscape 2085a QDS Board
Board: LS2085E-QDS, Board Arch: V1, Board version: B, boot from vBank: 4

Signed-off-by: Haikun Wang <haikun.wang@freescale.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
9 years agokeystone2: net: add mcast function to keyston2 Ethernet driver
Vitaly Andrianov [Wed, 8 Jul 2015 15:56:01 +0000 (11:56 -0400)]
keystone2: net: add mcast function to keyston2 Ethernet driver

The MCAST_TFTP support requires that network drivers has mcast functon
implemented. This commit adds dummy keystone2_eth_bcast_addr() to meet
the requirement. As far as the driver doesn't use ALE and doesn't filter
any incoming packets, the function is empty.

Signed-off-by: Vitaly Andrianov <vitalya@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agokeystone2: add wfi in to the core_spin loop
Vitaly Andrianov [Wed, 8 Jul 2015 15:40:14 +0000 (11:40 -0400)]
keystone2: add wfi in to the core_spin loop

When core A turning of core B, via tetris DPSC it places the core
B DPSC into transitional state. The core B has to execute wfi instruction
to move its DPSC to the OFF state. This patch add such instruction.

Signed-off-by: Vitaly Andrianov <vitalya@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
9 years agomisc: led: pca9551_led: Fix problem with multiple blink frequencies
Stefan Roese [Mon, 6 Jul 2015 11:35:55 +0000 (13:35 +0200)]
misc: led: pca9551_led: Fix problem with multiple blink frequencies

Only 2 frequencies are supported. The current driver implementation does
not always use the 2 last configured blink frequencies. This patch
fixes this problem. So that the last two entered frequencies are
active.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Tom Rini <trini@konsulko.com>
9 years agostm32f429-discovery: Use ttyS0 as the console device
rev13@wp.pl [Sun, 5 Jul 2015 10:54:23 +0000 (12:54 +0200)]
stm32f429-discovery: Use ttyS0 as the console device

Mainline kernel will be using this device name as well.

Signed-off-by: Kamil Lulko <rev13@wp.pl>
9 years agokbuild: use relative path more to include Makefile
Masahiro Yamada [Sat, 4 Jul 2015 16:56:57 +0000 (01:56 +0900)]
kbuild: use relative path more to include Makefile

Prior to this commit, it was impossible to use relative path to
include Makefiles from the top level Makefile because the option
"--include-dir=$(srctree)" becomes effective when Make enters into
sub Makefiles.

To use relative path in any places, this commit moves the option
above the "sub-make" target.

This was imported from Linux (commit a436bb7b8063) and adjusted for
U-Boot.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
9 years agokbuild: do not add $((generic-)offsets-file) to targets
Masahiro Yamada [Sat, 4 Jul 2015 16:56:56 +0000 (01:56 +0900)]
kbuild: do not add $((generic-)offsets-file) to targets

$(always) is added to targets by scripts/Makefile.build.
Moreover, filechk does not need .*.cmd files.

Adding these two files to targets is redundant.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agokbuild: sync with Linux 4.1
Masahiro Yamada [Sat, 4 Jul 2015 16:56:55 +0000 (01:56 +0900)]
kbuild: sync with Linux 4.1

Update some build scripts to match Linux 4.1.  Commit-based syncing
was done so as not to break U-Boot specific changes.
The previous big sync was from Linux 3.18-rc1 by commit 176d09827725
(kbuild: sync misc scripts with Linux 3.18-rc1).

The commits imported from Linux (some with adjustments) are:

[1] commit 9fb5e5372208973984a23ee6f5f025c05d364633
    Author: Robert Richter <rrichter@cavium.com>
    dts, kbuild: Factor out dtbs install rules to Makefile.dtbinst

[2] commit 371fdc77af44f4cb32475fd499e1d912ccc30890
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: collect shorthands into scripts/Kbuild.include

[3] commit a29b82326ed4eb5567b03c85b52c6891578d5a03
    Author: Michal Marek <mmarek@suse.cz>
    kbuild: Remove duplicate $(cmd) definition in Makefile.clean

[4] commit 1846dfbde3e8a53f3673dcb1c1b79fd9b3f8d40d
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: remove redundant -rR flag of hdr-inst

[5] commit 34948e0bbf98640fc1821751b01d2f0cd17d84d5
    Author: Michal Marek <mmarek@suse.cz>
    kbuild: Drop support for clean-rule

[6] commit a16c5f99a28c9945165c46da27fff8e6f26f8736
    Author: Michal Marek <mmarek@suse.cz>
    kbuild: Fix removal of the debian/ directory

[7] commit d0d38cd9e853db11e0242b3df4c9c3c4a663fbb4
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: use mixed-targets when two or more config targets are given

[8] commit dd33c03b18b3f2db791eb6a17c37d2de66e4de18
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: fix cc-ifversion macro

[9] commit 665d92e38f65d70796aad2b8e49e42e80815d4a4
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: do not add $(call ...) to invoke cc-version or cc-fullversion

[10] commit 6dcb4e5edf39e3b65a75ca76f087b2fdbee8a808
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: allow cc-ifversion to have the argument for false condition

[11] commit c0a80c0c27e5e65b180a25e6c4c2f7ef9e386cd3
    Author: Heiko Carstens <heiko.carstens@de.ibm.com>
    ftrace: allow architectures to specify ftrace compile options

[12] commit 0b24becc810dc3be6e3f94103a866f214c282394
    Author: Andrey Ryabinin <a.ryabinin@samsung.com>
    kasan: add kernel address sanitizer infrastructure

[13] commit 4218affdf57f938c04e3a916a9685ee27079f377
    Author: Masahiro Yamada <yamada.m@jp.panasonic.com>
    kbuild: remove warning about "make depend"

[14] commit 77479b38e2f58890eb221a0418357502a5b41cd6
    Author: Nathan Rossi <nathan.rossi@xilinx.com>
    kbuild: Create directory for target DTB

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Robert Richter <rrichter@cavium.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrey Ryabinin <a.ryabinin@samsung.com>
Signed-off-by: Nathan Rossi <nathan.rossi@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agokconfig: sync with Linux 4.1
Masahiro Yamada [Sat, 4 Jul 2015 16:56:54 +0000 (01:56 +0900)]
kconfig: sync with Linux 4.1

Update the files under scripts/kconfig/ to match Linux 4.1.
Some Kconfig sources have diverged from those in the kernel,
so commit-base syncing was done not to lose U-Boot specific updates.

The commits cherry-picked from Linux are:

[1] commit be8af2d54a66911693eddc556e4f7a866670082b
    Author: Bjørn Forsman <bjorn.forsman@gmail.com>
    kconfig/lxdialog: get ncurses CFLAGS with pkg-config

[2] commit 3943f42c11896ce82ad3da132c8a5630313bdd0e
    Author: Andrey Utkin <andrey.krieger.utkin@gmail.com>
    Replace mentions of "list_struct" to "list_head"

[3] commit e4e458b45c5861808674eebfea94cee2258bb2ea
    Author: Arjun Sreedharan <arjun024@gmail.com>
    calloc/xcalloc: Fix argument order

[4] commit 09950bc256e3628d275f90e016e6f5a039fbdcab
    Author: Olof Johansson <olof@lixom.net>
    merge_config.sh: Display usage if given too few arguments

[5] commit b6a2ab2cd4739a9573ed41677e53171987b8da34
    Author: Colin Ian King <colin.king@canonical.com>
    kconfig: use va_end to match corresponding va_start

[6] commit 70529b1a1784503169416df19ce3d68746401340
    Author: Michal Marek <mmarek@suse.cz>
    kconfig: Get rid of the P() macro in headers

[7] commit 463157444e377bf9b279101b1f16a94c4648c03a
    Author: Michal Marek <mmarek@suse.cz>
    kconfig: Remove dead code

[8] commit ad8d40cda3ad22ad9e8863d55a5c88f85c0173f0
    Author: Michal Marek <mmarek@suse.cz>
    kconfig: Remove unnecessary prototypes from headers

[9] commit de4619937229378e81f95e99c9866acc8e207d34
    Author: Masahiro Yamada <yamada.masahiro@socionext.com>
    kbuild: mergeconfig: fix "jobserver unavailable" warning

[10] commit b9fe99c5b994c6ddc57780993966b18899526c0b
    Author: Masahiro Yamada <yamada.masahiro@socionext.com>
    kbuild: mergeconfig: move an error check to merge_config.sh

[11] commit 371cfd4ff0611d8bc5d18bbb9cc6a2bc3d56cd3d
    Author: Masahiro Yamada <yamada.masahiro@socionext.com>
    kbuild: mergeconfig: remove redundant $(objtree)

[12] commit 3a975b8cfcbe026b535f83bde9a3c009bae214f9
    Author: Masahiro Yamada <yamada.masahiro@socionext.com>
    merge_config.sh: improve indentation

[13] commit bc8f8f5fc47cd02c2c5f3580dac2fe6695af1edd
    Author: Masahiro Yamada <yamada.masahiro@socionext.com>
    merge_config.sh: rename MAKE to RUNMAKE

[14] commit 63a91033d52e64a22e571fe84924c0b7f21c280d
    Author: Masahiro Yamada <yamada.masahiro@socionext.com>
    kbuild: add generic mergeconfig target, %.config

[15] commit 1cba0c305758c3c1786ecaceb03e142c95a4edc9
    Author: Michal Marek <mmarek@suse.cz>
    kconfig: Simplify Makefile

[16] commit 0a1f00a1c86421cc07cec87011c7cf4df68ee54b
    Author: Michal Marek <mmarek@suse.cz>
    kconfig: Do not print status messages in make -s mode

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Bjørn Forsman <bjorn.forsman@gmail.com>
Signed-off-by: Andrey Utkin <andrey.krieger.utkin@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Arjun Sreedharan <arjun024@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
9 years agoconfig: ti_omap5_common: Palmas power support in SPL
Paul Kocialkowski [Wed, 1 Jul 2015 22:20:25 +0000 (00:20 +0200)]
config: ti_omap5_common: Palmas power support in SPL

Palmas power support is required for OMAP5 devices such as the OMAP5 uEVM, that
need to e.g. enable MMC power at SPL stage.

This is especially important when booting from a peripheral (such as USB, UART),
where the bootrom will not enable power for the MMC device that will hold the
main U-Boot.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Reviewed-by: Tom Rini <trini@konsulko.com>
9 years agoMerge branch 'master' of http://git.denx.de/u-boot-sunxi
Tom Rini [Sat, 25 Jul 2015 13:04:18 +0000 (09:04 -0400)]
Merge branch 'master' of git.denx.de/u-boot-sunxi

9 years agosunxi: Enable both ehci and otg in host mode on various boards
Hans de Goede [Thu, 18 Jun 2015 16:19:22 +0000 (18:19 +0200)]
sunxi: Enable both ehci and otg in host mode on various boards

Now that the device-model port of the musb mode makes it possible, enable
both the ehci and otg in host mode on boards where the musb is wired up in
host only mode, either via an usb-a receptacle or via an usb <-> sata
converter.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: ga10h: Enable both otg and regular usb host controllers
Hans de Goede [Wed, 17 Jun 2015 19:16:59 +0000 (21:16 +0200)]
sunxi: ga10h: Enable both otg and regular usb host controllers

This allows using devices plugged into both ports of the tablet.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: musb: Stop treating not having a vbus-det gpio as an error
Hans de Goede [Thu, 18 Jun 2015 16:21:33 +0000 (18:21 +0200)]
sunxi: musb: Stop treating not having a vbus-det gpio as an error

On some boards the otg is wired up in host-only mode in this case we
have no vbus-det gpio.

Stop logging an error from sunxi_usb_phy_vbus_detect() in this case, and
stop treating sunxi_usb_phy_vbus_detect() returning a negative errno, as
if a charger is plugged into the otg port.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: musb: Improve output during probing
Hans de Goede [Thu, 18 Jun 2015 20:45:34 +0000 (22:45 +0200)]
sunxi: musb: Improve output during probing

When we return an error the usb core will print an error-message, so in this
case do not print anything.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
9 years agosunxi: musb: Use device-model for musb host mode
Hans de Goede [Wed, 17 Jun 2015 15:44:58 +0000 (17:44 +0200)]
sunxi: musb: Use device-model for musb host mode

Modify the sunxi musb glue to use the device-model for musb host mode.

This allows using musb in host mode together with other host drivers
such as ehci / ohci, which is esp. useful on boards which use the
musb controller in host-only mode, these boards have e.g. an usb-a
receptacle or an usb to sata converter attached to the musb controller.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: musb: Move musb config and platdata to the sunxi-musb glue
Hans de Goede [Wed, 17 Jun 2015 13:49:26 +0000 (15:49 +0200)]
sunxi: musb: Move musb config and platdata to the sunxi-musb glue

Move the musb config and platdata to the sunxi-musb glue, which is where
it really belongs. This is preparation patch for adding device-model
support for the sunxi-musb-host code.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: musb: Add id pin support
Hans de Goede [Sun, 14 Jun 2015 15:40:37 +0000 (17:40 +0200)]
sunxi: musb: Add id pin support

When in host mode check if there is a host cable inserted into the otg
port by checking the id pin. If there is no host cable return an error to
make usb_lowlevel_init() exit early, rather then waiting for 1 second
for a device which will never show up.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: musb: Move vbus check to sunxi_musb_enable
Hans de Goede [Sun, 14 Jun 2015 14:48:56 +0000 (16:48 +0200)]
sunxi: musb: Move vbus check to sunxi_musb_enable

This way it can be re-checked on "usb reset".

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: usb-phy: Add support for reading otg id pin value
Hans de Goede [Sun, 14 Jun 2015 15:29:53 +0000 (17:29 +0200)]
sunxi: usb-phy: Add support for reading otg id pin value

Add support for reading the id pin value of the otg connector to the usb
phy code.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agosunxi: Enable CMD_USB and USB_STORAGE by default on sunxi
Hans de Goede [Wed, 17 Jun 2015 18:54:07 +0000 (20:54 +0200)]
sunxi: Enable CMD_USB and USB_STORAGE by default on sunxi

Start using the new Kconfig options which are available for these now,
and simply always enable them by selecting them as sunxi builds always
include USB support.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Reviewed-by: Simon Glass <sjg@chromium.org>
9 years agosunxi: Update selects in arch/arm/Kconfig for DM conversions
Tom Rini [Tue, 30 Jun 2015 20:51:15 +0000 (16:51 -0400)]
sunxi: Update selects in arch/arm/Kconfig for DM conversions

With certain features being convert to DM now we want sunxi to default
to having DM enabled for ETH/SERIAL and USB in some cases.

Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Tom Rini <trini@konsulko.com>
[hdegoede@redhat.com: Also select CONFIG_USB for all sunxi builds]
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Reviewed-by: Simon Glass <sjg@chromium.org>
9 years agosunxi: Remove bogus uart entry from utoo-p66 dts file
Hans de Goede [Sat, 25 Jul 2015 09:20:38 +0000 (11:20 +0200)]
sunxi: Remove bogus uart entry from utoo-p66 dts file

At one point in time the utoo-p66 dts file in the kernel had a bogus
uart entry, and it seems like we synced with the kernel at just the wrong
moment.

This commit removes the bogus uart entry, which breaks booting the utoo-p66
when DM_SERIAL=y.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agoMerge git://git.denx.de/u-boot-usb
Tom Rini [Fri, 24 Jul 2015 20:39:56 +0000 (16:39 -0400)]
Merge git://git.denx.de/u-boot-usb

9 years agousb: ci_udc: fix request allocation when endpoints are disabled
Rob Herring [Fri, 24 Jul 2015 15:14:21 +0000 (10:14 -0500)]
usb: ci_udc: fix request allocation when endpoints are disabled

The ci_udc driver request allocation assumes that the endpoint descriptor
pointer is set to retrieve the endpoint number, but that is only true
when the endpoint is enabled. This results in a NULL ptr dereference
which for me happens to return 0 value. This causes the EP0 request
struct to be returned for other endpoints. Some gadget drivers like
fastboot and USB MS work fine, but ethernet does not.

Really, the ci_udc driver is the oddball here doing this EP0 special
case handling Stephen added. All the other drivers alloc/free functions
are pretty much the same with the only variation being the size of the
private struct. This could all be consolidated to a common function.

Signed-off-by: Rob Herring <robh@kernel.org>
Cc: Marek Vasut <marex@denx.de>
Acked-by: Stephen Warren <swarren@nvidia.com>
9 years agosunxi: spl: Remove redundant check from `board_mmc_init` for signature
Daniel Kochmański [Fri, 29 May 2015 15:21:00 +0000 (17:21 +0200)]
sunxi: spl: Remove redundant check from `board_mmc_init` for signature

Remove the boot signature check from board_mmc_init() in spl mode, as it
is already done in spl_boot_device() in this case, and update the comments
to reflect this.

Signed-off-by: Daniel Kochmański <dkochmanski@turtle-solutions.eu>
CC: Roy Spliet <r.spliet@ultimaker.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
[hdegoede@redhat.com: Disable the check only for SPL instead of always]
Acked-by: Hans De Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: spl: Detect at runtime where SPL was read from
Daniel Kochmański [Fri, 29 May 2015 14:55:42 +0000 (16:55 +0200)]
sunxi: spl: Detect at runtime where SPL was read from

Make possible using a single `u-boot-sunxi-with-spl.bin` binary for both NAND
memory and SD card. Detection where SPL was read from is implemented in
`spl_boot_device`.

Signed-off-by: Daniel Kochmański <dkochmanski@turtle-solutions.eu>
CC: Roy Spliet <r.spliet@ultimaker.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
[hdegoede@redhat.com: Some small coding style fixes]
Acked-by: Hans De Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agosunxi: Create helper function veryfing valid boot signature on MMC
Daniel Kochmański [Fri, 29 May 2015 14:55:41 +0000 (16:55 +0200)]
sunxi: Create helper function veryfing valid boot signature on MMC

This patch extracts checking for valid SD card "eGON.BT0" signature from
`board_mmc_init` into function `sunxi_mmc_has_egon_boot_signature`.

Buffer for mmc sector is allocated and freed at runtime. `panic` is
triggered on malloc failure.

Signed-off-by: Daniel Kochmański <dkochmanski@turtle-solutions.eu>
CC: Roy Spliet <r.spliet@ultimaker.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
[hdegoede@redhat.com: Small bugfix to make it work for devs other then mmc0]
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agommc: Protect `mmc_initialize` from initialising mmc multiple times
Daniel Kochmański [Fri, 29 May 2015 14:55:43 +0000 (16:55 +0200)]
mmc: Protect `mmc_initialize` from initialising mmc multiple times

`mmc_initialize` might be called multiple times leading to the mmc-controllers
being initialised twice, and initialising the `mmc_devices` list head twice
which may lead to memory leaks.

Signed-off-by: Daniel Kochmański <dkochmanski@turtle-solutions.eu>
CC: Roy Spliet <r.spliet@ultimaker.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
CC: Pantelis Antoniou <panto@antoniou-consulting.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agoarm: marvell: Increase MAXARGS to 32
Stefan Roese [Thu, 23 Jul 2015 08:26:11 +0000 (10:26 +0200)]
arm: marvell: Increase MAXARGS to 32

This makes more complex U-Boot scripts possible.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
9 years agoarm: mvebu: db-88f6820-gp: Add SDIO/MMC SPL boot support
Stefan Roese [Mon, 20 Jul 2015 09:20:40 +0000 (11:20 +0200)]
arm: mvebu: db-88f6820-gp: Add SDIO/MMC SPL boot support

This patch adds the configuration options to boot via SDIO/MMC on the
Marvell DB-88F6820-GP Armada A38x board. The default boot device
is still SPI NOR flash.

To enable MMC booting on this board 2 things need to be changes:
a) Change kwbimage.cfg
   BOOT_FROM   sdio
b) In the config header select
   #define CONFIG_SPL_BOOT_DEVICE SPL_BOOT_SDIO_MMC_CARD

The generated image needs to be copied to the first bootable MMC
partition:

dd if=u-boot-spl.kwb of=/dev/sdX1

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Dirk Eibach <eibach@gdsys.de>
9 years agospl: spl_mmc: Add option to boot from a MMC partition with offset
Stefan Roese [Mon, 20 Jul 2015 09:20:39 +0000 (11:20 +0200)]
spl: spl_mmc: Add option to boot from a MMC partition with offset

This patch introduces the option to boot from a MMC card parition with
an offset. This can be done by using both defines together:

define CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_PARTITION 1
define CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR    ((160 << 10) / 512)

The example above loads the main U-Boot at offset 160KiB from the MMC
partition 1.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Dirk Eibach <eibach@gdsys.de>
Cc: Tom Rini <trini@konsulko.com>
9 years agokwbimage: Rename CONFIG_SYS_SPI_U_BOOT_OFFS to CONFIG_SYS_U_BOOT_OFFS
Stefan Roese [Mon, 20 Jul 2015 09:20:38 +0000 (11:20 +0200)]
kwbimage: Rename CONFIG_SYS_SPI_U_BOOT_OFFS to CONFIG_SYS_U_BOOT_OFFS

To use this offset for other boot device (like SDIO/MMC), lets rename
it to a more generic name. This will be used be the SDIO/MMC SPL boot
support for the A38x.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Dirk Eibach <eibach@gdsys.de>
9 years agokwbimage: Add support for SDIO/MMC boot device selection
Stefan Roese [Mon, 20 Jul 2015 09:20:37 +0000 (11:20 +0200)]
kwbimage: Add support for SDIO/MMC boot device selection

This patch adds support to select the "sdio" as boot device in the
kwbimage.cfg file. This line selects this SDIO device:

BOOT_FROM  sdio

Tested on Marvell DB-88F6820-GP board.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Dirk Eibach <eibach@gdsys.de>
9 years agoarm: mvebu: Add SPL SDIO/MMC boot support
Stefan Roese [Mon, 20 Jul 2015 09:20:36 +0000 (11:20 +0200)]
arm: mvebu: Add SPL SDIO/MMC boot support

This patch adds basic SDIO/MMC booting support to MVEBU SoC's. Since
I don't know of a way to test the boot-device upon runtime, this patch
hardcodes the spl_boot_device instead.

Tested on Marvell DB-88F6820-GP board.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Dirk Eibach <eibach@gdsys.de>
9 years agoMerge git://git.denx.de/u-boot-uniphier
Tom Rini [Thu, 23 Jul 2015 15:46:05 +0000 (11:46 -0400)]
Merge git://git.denx.de/u-boot-uniphier

9 years agoARM: UniPhier: document reference support card
Masahiro Yamada [Tue, 21 Jul 2015 05:04:24 +0000 (14:04 +0900)]
ARM: UniPhier: document reference support card

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: dts: UniPhier: add I2C ch4 device node for PH1-sLD3
Masahiro Yamada [Tue, 21 Jul 2015 05:04:23 +0000 (14:04 +0900)]
ARM: dts: UniPhier: add I2C ch4 device node for PH1-sLD3

This I2C device is used SoC-internally for controlling the DMD core.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: UniPhier: add PH1-sLD3 SoC support
Masahiro Yamada [Tue, 21 Jul 2015 05:04:22 +0000 (14:04 +0900)]
ARM: UniPhier: add PH1-sLD3 SoC support

The init code for UMC (Unified Memory Controller) and PLL has not
been mainlined yet, but U-boot proper should work.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: UniPhier: update CONFIG_BOOTARGS
Masahiro Yamada [Tue, 21 Jul 2015 05:04:21 +0000 (14:04 +0900)]
ARM: UniPhier: update CONFIG_BOOTARGS

Both "earlyprintk" and "loglevel=8" should be useful for the
development.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: UniPhier: disable CONFIG_FIT
Masahiro Yamada [Tue, 21 Jul 2015 05:04:20 +0000 (14:04 +0900)]
ARM: UniPhier: disable CONFIG_FIT

To use FIT boot, we have to describe Image Tree Source in addition.
So, it is not intended for beginners.  Disable it by default.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: UniPhier: add environment variable for TFTP boot
Masahiro Yamada [Tue, 21 Jul 2015 05:04:19 +0000 (14:04 +0900)]
ARM: UniPhier: add environment variable for TFTP boot

The command "run tftpboot" downloads some files onto the RAM
via TFTP and boots the kernel.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: UniPhier: add environment sets for non-FIT Linux boot
Masahiro Yamada [Tue, 21 Jul 2015 05:04:18 +0000 (14:04 +0900)]
ARM: UniPhier: add environment sets for non-FIT Linux boot

Currently, the environment variables "norboot" and "nandboot" only
work with CONFIG_FIT, but we do not want to depend on CONFIG_FIT to
boot the kernel.

This commit adds environments useful for booting Linux with separate
uImage + ramdisk + DTB.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoARM: dts: UniPhier: add device-specific compatible string for EEPROM
Masahiro Yamada [Mon, 13 Jul 2015 01:35:34 +0000 (10:35 +0900)]
ARM: dts: UniPhier: add device-specific compatible string for EEPROM

For the record, describe exactly which device of which vendor
is used on this board.

I2C EEPROM is bound by the generic compatible string, "i2c-eeprom",
so this commit has no impact on the functionality.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoMerge git://git.denx.de/u-boot-marvell
Tom Rini [Thu, 23 Jul 2015 13:02:28 +0000 (09:02 -0400)]
Merge git://git.denx.de/u-boot-marvell

9 years agoarm: mvebu: db-88f6820: Add SPL support with DDR init code
Stefan Roese [Wed, 25 Mar 2015 12:35:15 +0000 (13:35 +0100)]
arm: mvebu: db-88f6820: Add SPL support with DDR init code

This patch adds SPL support for the Marvell DB-88F6820-GP board.
With this change, the bin_hdr from the original Marvell U-boot
is not needed any more on this board. The sources from bin_hdr
(SERDES/PHY and DDR setup) are now integrated in mainline
U-Boot. And this patch enables them for this board.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: a38x: Use correct PEX register access macros
Stefan Roese [Mon, 8 Jun 2015 15:01:26 +0000 (17:01 +0200)]
arm: mvebu: a38x: Use correct PEX register access macros

Remove the incorrect PEX macros from the DDR header. And insert the
correct ones in ctrl_pex.h instead.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: Add Armada 38x DDR3 training code from Marvell bin_hdr
Stefan Roese [Thu, 26 Mar 2015 14:36:56 +0000 (15:36 +0100)]
arm: mvebu: Add Armada 38x DDR3 training code from Marvell bin_hdr

This patch adds the DDR3 setup and training code taken from the Marvell
U-Boot repository. This code used to be included as a binary (bin_hdr)
into the Armada A38x boot image. Not linked with the main U-Boot. With this
code addition and the serdes/PHY setup code, the Armada A38x support
in mainline U-Boot is finally self-contained. So the complete image
for booting can be built from mainline U-Boot. Without any additional
external inclusion.

Note:
This code has undergone many hours (days!) of coding-style cleanup and
refactoring. It still is not checkpatch clean though, I'm afraid. As the
factoring of the code has so many levels of indentation that many lines
are longer than 80 chars.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: drivers/ddr: Move Armada XP DDR init code into new directory
Stefan Roese [Wed, 25 Mar 2015 11:51:18 +0000 (12:51 +0100)]
arm: mvebu: drivers/ddr: Move Armada XP DDR init code into new directory

With the upcoming addition of the Armada 38x DDR support, which is not
compatible to the Armada XP DDR init code, we need to introduce a new
directory infrastructure. To support multiple Marvell DDR controller.

This will be the new structure:

     drivers/ddr/marvell/axp
     Supporting Armada XP (AXP) devices (and perhaps Armada 370)

     drivers/ddr/marvell/a38x
     Supporting Armada 38x devices (and perhaps Armada 39x)

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: Add Armada 38x SERDES / PHY init code from Marvell bin_hdr
Stefan Roese [Mon, 20 Apr 2015 07:31:27 +0000 (09:31 +0200)]
arm: mvebu: Add Armada 38x SERDES / PHY init code from Marvell bin_hdr

This code is ported from the Marvell bin_hdr code into mainline
SPL U-Boot. It needs to be executed very early so that the devices
connected to the serdes PHY are configured correctly.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: serdes: Move Armada XP SERDES / PHY init code into new directory
Stefan Roese [Mon, 20 Apr 2015 07:28:12 +0000 (09:28 +0200)]
arm: mvebu: serdes: Move Armada XP SERDES / PHY init code into new directory

With the upcoming addition of the Armada 38x SPL support, which is not
compatible to the Armada XP SERDES init code, we need to introduce a new
directory infrastructure. So lets move the AXP serdes init code into
a new directory. This way the A38x code can be added in a clean way.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoMakefile: Fix mvebu build target to use SPL load and exe-address
Stefan Roese [Mon, 20 Apr 2015 09:17:16 +0000 (11:17 +0200)]
Makefile: Fix mvebu build target to use SPL load and exe-address

The u-boot-spl.kwb build target needs the SPL text-base
(CONFIG_SPL_TEXT_BASE) as load and execution address.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: Disable MMU before changing register base address
Stefan Roese [Fri, 24 Apr 2015 08:49:11 +0000 (10:49 +0200)]
arm: mvebu: Disable MMU before changing register base address

Only with disabled MMU its possible to switch the base register address on
Armada 38x. Without this the SDRAM located at >= 0x4000.0000 is also not
accessible, as its still locked to cache.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: spl.c: Add call to board_early_init_f()
Stefan Roese [Fri, 17 Apr 2015 16:13:06 +0000 (18:13 +0200)]
arm: mvebu: spl.c: Add call to board_early_init_f()

Pin muxing needs to be done before UART output, since on A38x the UART
pins need some re-muxing for output to work.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoarm: mvebu: Use default reg base address for SPL on A38x
Stefan Roese [Fri, 17 Apr 2015 16:12:41 +0000 (18:12 +0200)]
arm: mvebu: Use default reg base address for SPL on A38x

On A38x switching the regs base address without running from
SDRAM doesn't seem to work. So let the SPL still use the
default base address and switch to the new address in the
mail u-boot later.

Signed-off-by: Stefan Roese <sr@denx.de>
9 years agoci_udc: fix 64-bit compile warnings
Stephen Warren [Wed, 22 Jul 2015 21:16:20 +0000 (15:16 -0600)]
ci_udc: fix 64-bit compile warnings

This is the same as f72d8320b605 "usb: ci_udc: fix warnings on 64-bit
builds", but more.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
9 years agoarm: mvebu: Call timer_init early before PHY and DDR init
Stefan Roese [Wed, 15 Jul 2015 13:36:52 +0000 (15:36 +0200)]
arm: mvebu: Call timer_init early before PHY and DDR init

Without calling timer_init(), the xdelay() functions return immediately.
We need to call timer_init() early, so that these functions work and
the PHY and DDR init code works correctly.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Anton Schubert <anton.schubert@gmx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
9 years agoarm: mvebu: add Armada XP SATA support
Anton Schubert [Wed, 15 Jul 2015 12:50:05 +0000 (14:50 +0200)]
arm: mvebu: add Armada XP SATA support

This patch initializes the SATA address windows on Armada XP and
allows it to work with the existing mvsata_ide driver.
It also adds the necessary configuration for the db-mv784mp-gp board.

Signed-off-by: Anton Schubert <anton.schubert@gmx.de>
Tested-by: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
9 years agolinux_compat: handle __GFP_ZERO in kmalloc()
Masahiro Yamada [Mon, 13 Jul 2015 04:17:07 +0000 (13:17 +0900)]
linux_compat: handle __GFP_ZERO in kmalloc()

Currently, kzalloc() returns zero-filled memory, while kmalloc()
simply ignores the second argument and never fills the memory
area with zeros.

I want kmalloc(size, __GFP_ZERO) to behave as kzalloc() does,
which will make it easier to add more memory allocator variants.

With the introduction of __GFP_ZERO flag, going forward, kzmalloc()
variants can fall back to kmalloc() enabling the __GFP_ZERO flag.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Acked-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Lukasz Majewski <l.majewski@samsung.com>
9 years agolinux_compat: move vzalloc() to header file as an inline function
Masahiro Yamada [Mon, 13 Jul 2015 04:17:06 +0000 (13:17 +0900)]
linux_compat: move vzalloc() to header file as an inline function

The vzalloc(size) is equivalent to kzalloc(size, 0).  Move it to
include/linux/compat.h as an inline function in order to avoid the
function call overhead.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agolinux_compat: remove cpu_relax() define
Masahiro Yamada [Mon, 13 Jul 2015 04:17:05 +0000 (13:17 +0900)]
linux_compat: remove cpu_relax() define

The macro cpu_relax() is defined by several headers in different
ways.

arch/{arm,avr32,mips}/include/asm/processor.h defines it as follows:
  #define cpu_relax() barrier()

On the other hand, include/linux/compat.h defines it as follows:
  #define cpu_relax() do {} while (0)

If both headers are included from the same source file, the warning
  warning: "cpu_relax" redefined [enabled by default]
is displayed.

It effectively makes it impossible to include <linux/compat.h>
from some sources.  Drop the latter.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Acked-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Lukasz Majewski <l.majewski@samsung.com>
9 years agox86: delete unneeded declarations of disable_irq() and enable_irq()
Masahiro Yamada [Mon, 13 Jul 2015 04:17:04 +0000 (13:17 +0900)]
x86: delete unneeded declarations of disable_irq() and enable_irq()

These two declarations in arch/x86/include/asm/interrupt.h conflict
with ones in include/linux/compat.h, so x86 boards cannot include
<linux/compat.h>.

The comment /* arch/x86/lib/interrupts.c */ is bogus now, and we do
not see any definitions of disable_irq() and enable_irq() in there.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agousb: ehci-marvell: Drop wrl accessor function
Stefan Roese [Wed, 22 Jul 2015 08:01:30 +0000 (10:01 +0200)]
usb: ehci-marvell: Drop wrl accessor function

This patch removes the wrl accessor function from the Marvell EHCI
driver by replacing it with the writel function.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
9 years agousb: CONFIG_USB_FASTBOOT prefix replacement for consistency
Paul Kocialkowski [Mon, 20 Jul 2015 10:38:22 +0000 (12:38 +0200)]
usb: CONFIG_USB_FASTBOOT prefix replacement for consistency

FASTBOOT is defined both by CONFIG_USB_FUNCTION_FASTBOOT AND CONFIG_CMD_FASTBOOT, so it doesn't
make much sense to have a CONFIG_USB_FASTBOOT prefix for fastboot-specific options, especially
given that other config options for fastboot use the CONFIG_FASTBOOT prefix.

This replaces the CONFIG_USB_FASTBOOT prefix with CONFIG_FASTBOOT, for consistency.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agoimplement Fastboot via USB OTG on bcm28155_ap boards
Jiandong Zheng [Thu, 9 Jul 2015 21:26:40 +0000 (14:26 -0700)]
implement Fastboot via USB OTG on bcm28155_ap boards

Signed-off-by: Jiandong Zheng <jdzheng@broadcom.com>
Signed-off-by: Steve Rae <srae@broadcom.com>
9 years agousb: gadget: bcm_udc_otg files
Jiandong Zheng [Thu, 9 Jul 2015 21:26:39 +0000 (14:26 -0700)]
usb: gadget: bcm_udc_otg files

Add the required files for the Broadcom UDC OTG interface.

Signed-off-by: Jiandong Zheng <jdzheng@broadcom.com>
Signed-off-by: Steve Rae <srae@broadcom.com>
9 years agog_dnl: add missing declaration
Steve Rae [Thu, 9 Jul 2015 21:26:38 +0000 (14:26 -0700)]
g_dnl: add missing declaration

Signed-off-by: Steve Rae <srae@broadcom.com>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Test HW: Exynos 4210 (trats), Exynos 5422 (Odroid XU3)

9 years agousb: gadget: fastboot: Dequeue the previous IN request for the current request
Paul Kocialkowski [Sat, 4 Jul 2015 14:46:16 +0000 (16:46 +0200)]
usb: gadget: fastboot: Dequeue the previous IN request for the current request

Recent versions of the fastboot tool will query the partition type before doing
an operation on a partition (such as erase, flash, etc). It will then submit
the operation as soon as the response for the partition type is received.

Usually, the MUSB controller will see that the partition type request return
status was read by the host at the very same time as the actual operation
request is submitted by the host. However, the operation will be read first
(int_rx is handled first in musb_interrupt) and after it is completed, the
fastboot USB gadget driver will send another return status. Hence, this happens
before the musb gadget framework has had a chance to handle the previous
acknowledgement that the host read the return status and dequeue the request.

The host will then usually empty the FIFO by the time musb_interrupt gets around
handling the return status acknowledgement (for the previous request, this is
still on the same musb_interrupt call), so no other interrupt is generated and
the most recent return status acknowledgement remains unaccounted for.

It will then be used as a response for the next command, and the proper response
for it will be delayed to the next command, and so on.

Dequeuing the previous IN request in the fastboot code ensures that no previous
return status remains. It is acceptable to do it since there is no callback to
it anyways.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: gadget: fastboot: Request status and length check in rx handler
Paul Kocialkowski [Sat, 4 Jul 2015 14:46:15 +0000 (16:46 +0200)]
usb: gadget: fastboot: Request status and length check in rx handler

This avoids handling requests that have an error status or no data.
In particular, this avoids showing unnecessary error messages when the USB
gadget gets disconnected (e.g. with fastboot continue) and the fastboot USB
gadget driver sends an error back to the host (that has disconnected already).

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agofastboot: Dynamic controller index for usb_gadget_handle_interrupts
Paul Kocialkowski [Sun, 24 May 2015 09:41:04 +0000 (11:41 +0200)]
fastboot: Dynamic controller index for usb_gadget_handle_interrupts

Since we're now using a dynamic controller index for fastboot too,
usb_gadget_handle_interrupts should be using it instead of 0 (despite the fact
that it's currently not being used at all in the musb-new implementation).

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: gadget: Weak board_usb_init/cleanup definitions in USB download gadget code
Paul Kocialkowski [Fri, 12 Jun 2015 17:57:01 +0000 (19:57 +0200)]
usb: gadget: Weak board_usb_init/cleanup definitions in USB download gadget code

Weak versions of board_usb_init and board_usb_cleanup are defined in common USB
host code, but it is also used for USB device gadgets, so we also need a weak
definition of it when there is no USB host enabled.
Both weak definitions do not conflict.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Test HW: Odroid_XU3 (Exynos5422), trats (Exynos4210)

9 years agousb: board_usb_init and board_usb_cleanup calls in the fastboot command
Paul Kocialkowski [Fri, 12 Jun 2015 17:57:00 +0000 (19:57 +0200)]
usb: board_usb_init and board_usb_cleanup calls in the fastboot command

Each USB download function command calls board_usb_init before registering the
USB gadget and board_usb_cleanup after de-registering it. On devices currently
using fasboot, musb-new is usually initialized earlier, but some other boards
might need the board_usb_init call to properly initialize musb-new.

This requires adding an argument (the USB controller index) to the fastboot
command, as it is currently done with other USB download gadget functions.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Test HW: Odroid_XU3 (Exynos5422), trats (Exynos4210)

9 years agousb: Fastboot function config for better consistency with other functions
Paul Kocialkowski [Fri, 12 Jun 2015 17:56:59 +0000 (19:56 +0200)]
usb: Fastboot function config for better consistency with other functions

USB download gadget functions such as thor and dfu have a separate config option
for the USB gadget part of the code, independent from the command part.
This switches the fastboot USB gadget to the same scheme, for better
consistency.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Test HW: Odroid_XU3 (Exynos5422), trats (Exynos4210)

9 years agousb: USB download gadget and functions config options coherent naming
Paul Kocialkowski [Fri, 12 Jun 2015 17:56:58 +0000 (19:56 +0200)]
usb: USB download gadget and functions config options coherent naming

This introduces a coherent scheme for naming USB download gadget and functions
config options. The download USB gadget config option is moved to
CONFIG_USB_GADGET_DOWNLOAD for better consistency with other gadgets and each
function's config option is moved to a CONFIG_USB_FUNCTION_ prefix.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Test HW: Odroid_XU3 (Exynos5422), trats (Exynos4210)

9 years agodrivers: usb: fsl: Remove LS102XA immap inclusion
Nikhil Badola [Wed, 24 Jun 2015 05:22:48 +0000 (10:52 +0530)]
drivers: usb: fsl: Remove LS102XA immap inclusion

Remove LS102XA immap header inclusion from xhci fsl driver.
It removes redefinition warnings when built for platforms
other than LS102XA

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
9 years agoinclude: usb: Map USB controller base addresses for LS2085A
Nikhil Badola [Tue, 23 Jun 2015 03:48:20 +0000 (09:18 +0530)]
include: usb: Map USB controller base addresses for LS2085A

Map USB XHCI controller base addresses for LS2085A SOC

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
9 years agoinclude: usb: Move USB controller base address mapping
Nikhil Badola [Tue, 23 Jun 2015 03:48:03 +0000 (09:18 +0530)]
include: usb: Move USB controller base address mapping

Move USB controller Base address mapping from ls102xa immap
to fsl xhci header. This is required to remove any warnings when
controller base addresses are mapped for multiple platforms
in their respective files.

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
9 years agodrivers: usb: fsl: Implement Erratum A-009116 for XHCI controller
Nikhil Badola [Tue, 23 Jun 2015 03:47:49 +0000 (09:17 +0530)]
drivers: usb: fsl: Implement Erratum A-009116 for XHCI controller

This adjusts (micro)frame length to appropriate value thus
avoiding USB devices to time out over a longer run

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
9 years agodrivers: usb: fsl: Remove warnings for 64-bit architectures
Nikhil Badola [Tue, 23 Jun 2015 03:47:32 +0000 (09:17 +0530)]
drivers: usb: fsl: Remove warnings for 64-bit architectures

Replace uint32_t with uintptr_t to remove compilation
warnings for 64-bit architectures.

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
9 years agols1021aqds: Enable USB IP support
Ramneek Mehresh [Fri, 29 May 2015 09:17:22 +0000 (14:47 +0530)]
ls1021aqds: Enable USB IP support

Enable USB IP support for both EHCI and XHCI for
ls1021aqds platform

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agols1021atwr: Enable USB IP support
Ramneek Mehresh [Fri, 29 May 2015 09:17:21 +0000 (14:47 +0530)]
ls1021atwr: Enable USB IP support

Enable USB IP support for both EHCI and XHCI for
ls1021atwr platform

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agoarch: arm: fsl: Add XHCI support for LS1021A
Ramneek Mehresh [Fri, 29 May 2015 09:17:20 +0000 (14:47 +0530)]
arch: arm: fsl: Add XHCI support for LS1021A

Add base register address information for USB
XHCI controller on LS1021A

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agousb: fsl: Add XHCI driver support
Ramneek Mehresh [Fri, 29 May 2015 09:17:19 +0000 (14:47 +0530)]
usb: fsl: Add XHCI driver support

Add xhci driver support for all FSL socs

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agousb: xhci: keystone: Remove common dwc3 drv functions calls
Ramneek Mehresh [Fri, 29 May 2015 09:17:18 +0000 (14:47 +0530)]
usb: xhci: keystone: Remove common dwc3 drv functions calls

Remove all redundant dwc3 driver function calls that
are defined by dwc3 driver

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agousb: xhci: omap: Remove common dwc3 drv functions calls
Ramneek Mehresh [Fri, 29 May 2015 09:17:17 +0000 (14:47 +0530)]
usb: xhci: omap: Remove common dwc3 drv functions calls

Remove all redundant dwc3 driver function calls that
are defined by dwc3 driver

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agousb: xhci: exynos: Remove common dwc3 drv functions calls
Ramneek Mehresh [Fri, 29 May 2015 09:17:16 +0000 (14:47 +0530)]
usb: xhci: exynos: Remove common dwc3 drv functions calls

Remove all redundant dwc3 driver function calls that
are defined by dwc3 driver

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agousb: dwc3: Add DWC3 controller driver support
Ramneek Mehresh [Fri, 29 May 2015 09:17:15 +0000 (14:47 +0530)]
usb: dwc3: Add DWC3 controller driver support

Add support for DWC3 XHCI controller driver

Signed-off-by: Ramneek Mehresh <ramneek.mehresh@freescale.com>
9 years agonet: Allow drivers to return -ENOSYS with the write_hwaddr() method
Simon Glass [Mon, 6 Jul 2015 22:47:55 +0000 (16:47 -0600)]
net: Allow drivers to return -ENOSYS with the write_hwaddr() method

Some drivers may want to implement this method for some of their devices but
not for others. So it is not possible to just leave the operation out of
the table. Drivers could get around this by masquerading as two separate
drivers but that seems unpleasant.

Allow the driver to return an error when it does not want to process the
write_hwaddr() method.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: usb: eth: Add driver-model support to the asix driver
Simon Glass [Mon, 6 Jul 2015 22:47:54 +0000 (16:47 -0600)]
dm: usb: eth: Add driver-model support to the asix driver

This USB Ethernet driver is quite widely use. Allow it to work with
CONFIG_DM_ETH enabled. Most of the code remains common but there is a new
packet receive flow which is handled specially.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: usb: Add driver-model support to ehci-pci
Simon Glass [Mon, 6 Jul 2015 22:47:53 +0000 (16:47 -0600)]
dm: usb: Add driver-model support to ehci-pci

Support driver model in this driver. This uses the normal USB driver search
mechanism. Any EHCI controllers will be set up as they are found during
usb_init().

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Marek Vasut <marex@denx.de>
9 years agox86: minnowmax: Drop the cache line size hack
Simon Glass [Mon, 6 Jul 2015 22:47:52 +0000 (16:47 -0600)]
x86: minnowmax: Drop the cache line size hack

Now that the RTL8169 driver warning is fixed we can drop this. The incorrect
value is causing problems with USB EHCI.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agodm: usb: Adjust the USB_DEVICE() macro naming
Simon Glass [Mon, 6 Jul 2015 22:47:51 +0000 (16:47 -0600)]
dm: usb: Adjust the USB_DEVICE() macro naming

In Linux USB_DEVICE() is used to declare a USB device by vendor/device ID.
We should follow the same convention in U-Boot. Rename the existing
USB_DEVICE() macro to U_BOOT_USB_DEVICE() and bring in the USB_DEVICE()
macro from Linux for use in U-Boot.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: usb: eth: Support driver model with USB Ethernet
Simon Glass [Mon, 6 Jul 2015 22:47:50 +0000 (16:47 -0600)]
dm: usb: eth: Support driver model with USB Ethernet

At present USB Ethernet does not work with CONFIG_DM_ETH. Add driver model
support to this feature, so that it can work alongside other Ethernet
devices with driver model.

It was found that quite a bit of code is common in most of the USB Ethernet
drivers. Add this code to the common layer to reduce the amount of duplicate
code needed in USB Ethernet drivers when CONFIG_DM_ETH is used.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Marek Vasut <marex@denx.de>
9 years agodm: eth: Avoid blocking on packet reception
Simon Glass [Mon, 6 Jul 2015 22:47:49 +0000 (16:47 -0600)]
dm: eth: Avoid blocking on packet reception

Some devices can take a long time to work out whether they have a new packet
or now. For example the ASIX USB Ethernet dongle can take 5 seconds to do
this, since it waits until it gets a new packet on the wire before allowing
the USB bulk read packet to be submitted.

At present with driver mode the Ethernet receive code reads 32 packets. This
can take a very long time if we must wait for all 32 packets. The old code
(before driver model) worked by reading a single set of packets from the USB
device, then processing all the packets with in. It would be nice to use
the same behaviour with driver model.

Add a flag to the receive method which indicates that the driver should try
to find a packet if available, by consulting the hardware. When the flag is
not set, it should just return any packet data it has already received. If
there is none, it should return -EAGAIN so that the loop will terminate.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: usb: Avoid using USB ethernet with CONFIG_DM_USB and no DM_ETH
Simon Glass [Mon, 6 Jul 2015 22:47:48 +0000 (16:47 -0600)]
dm: usb: Avoid using USB ethernet with CONFIG_DM_USB and no DM_ETH

If driver model is used for Ethernet then USB Ethernet does not build. This
can be made to work with driver model is used for USB also. Add #ifdef logic
to make this clear when building.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Marek Vasut <marex@denx.de>
9 years agodm: usb: Correct the struct usb_driver_entry comment
Simon Glass [Mon, 6 Jul 2015 22:47:47 +0000 (16:47 -0600)]
dm: usb: Correct the struct usb_driver_entry comment

The function documentation is incorrect. Fix it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Marek Vasut <marex@denx.de>