karo-tx-uboot.git
2 years agoarm: remove bogus cp_delay() function karo-tx25 KARO-TX25-2017-08-17
Lothar Waßmann [Thu, 17 Aug 2017 10:37:11 +0000 (12:37 +0200)]
arm: remove bogus cp_delay() function

Due to a missing 'volatile' in the get_cr() function, the asm
instruction may be optimized away by the compiler.
This bug has been papered over with a delay loop after reading the
CP15 control register.
Fix the asm statement and remove the bogus workaround.

4 years agoti_armv7_common: env: Use partuuid for detecting mmc root fs
Lokesh Vutla [Fri, 28 Aug 2015 08:05:08 +0000 (13:35 +0530)]
ti_armv7_common: env: Use partuuid for detecting mmc root fs

Linux kernel can enumerate mmc sd as either mmcblk0 or mmcblk1.
But u-boot default environment assumes that sd always populates
as mmcblk0. With this the root fs is not being mounted when
mmc sd is enumerated as mmcblk1.
So use partuuid to update root= option in default environment.

Reported-by: Yan Liu <yan-liu@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
4 years agoti_armv7_common: env: Consolidate MMC args
Lokesh Vutla [Fri, 28 Aug 2015 08:05:07 +0000 (13:35 +0530)]
ti_armv7_common: env: Consolidate MMC args

Define default mmc args in ti_armv7_common.h so that all
TI platforms can reuse.

Reported-by: Yan Liu <yan-liu@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
4 years agoti_omap4_common: use ext4 fs as default.
Lokesh Vutla [Fri, 28 Aug 2015 08:05:06 +0000 (13:35 +0530)]
ti_omap4_common: use ext4 fs as default.

All TI SoCs expect filesystem to be ext4, omap4_common is the only one
with ext3. move omap4 to ext4 so that we can start consolidating MMC
arguments.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
4 years agoARM: DRA7: emif: Fix disabling/enabling of refreshes
Lokesh Vutla [Fri, 28 Aug 2015 06:58:25 +0000 (12:28 +0530)]
ARM: DRA7: emif: Fix disabling/enabling of refreshes

clrsetbits_le32/clrbits_le32 takes mask of the bits as input that
are needed to be set/clear. But emif driver passes the shift of the bits.
Fixing it here.

Reported-by: Mark Mckeown <m-mckeown@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
4 years agolpc32xx: remove duplicated DMA_CLK_ENABLE bit definition
Vladimir Zapolskiy [Thu, 27 Aug 2015 00:16:48 +0000 (03:16 +0300)]
lpc32xx: remove duplicated DMA_CLK_ENABLE bit definition

Because there is an originally defined CLK_DMA_ENABLE macro in clk.h,
no reason to add another DMA_CLK_ENABLE macro with the same value.

Remove DMA_CLK_ENABLE, since it does not follow naming convention from
the code, this implies renaming of DMA_CLK_ENABLE to CLK_DMA_ENABLE in
lpc32xx/devices.c file.

Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Tested-by: Sylvain Lemieux <slemieux@tycoint.com>
4 years agodriver: misc: correct Kconfig entry
Peng Fan [Wed, 26 Aug 2015 07:41:33 +0000 (15:41 +0800)]
driver: misc: correct Kconfig entry

Should use FSL_SEC_MON, not CONFIG_FSL_SEC_MON as Kconfig entry.

Signed-off-by: Peng Fan <Peng.Fan@freescale.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: York Sun <yorksun@freescale.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Tom Rini <trini@konsulko.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agoboard/BuR: simplify default IP-setup on B&R boards.
Hannes Petermaier [Tue, 25 Aug 2015 11:55:41 +0000 (13:55 +0200)]
board/BuR: simplify default IP-setup on B&R boards.

To simplify and having a common default IP-setup on all B&R boards we
introduce an environment variable "brdefaultip" which does following.

Test if ${ipaddr} is empty, if yes it set's up some defaults:
- ipaddr   : 192.168.60.1
- netmask  : 255.255.255.0
- gatewayip: 192.168.60.254
- serverip : 192.168.60.254

This environment is ran from CONFIG_PREBOOT.

All other "tricks" are dropped.

Signed-off-by: Hannes Schmelzer <oe5hpm@oevsv.at>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agofat: handle paths that include ../
Stephen Warren [Wed, 29 Jul 2015 03:55:03 +0000 (21:55 -0600)]
fat: handle paths that include ../

The FAT code contains a special case to parse the root directory. This
is needed since the root directory location/layout on disk is special
cased for FAT12/16. In particular, the location and size of the FAT12/16
root directory is hard-coded and contiguous, whereas all FAT12/16 non-root
directories, and all FAT32 directories, are stored in a non-contiguous
fashion, with the layout represented by a linked-list of clusters in the
FAT.

If a file path contains ../ (for example /extlinux/../bcm2835-rpi-cm.dtb),
it is possible to need to parse the root directory for the first element
in the path (requiring application of the special case), then a sub-
directory (in the general way), then re-parse the root directory (again
requiring the special case). However, the current code in U-Boot only
applies the special case for the very first path element, and never for
any later path element. When reparsing the root directory without
applying the special case, any file in a sector (or cluster?) other than
the first sector/cluster of the root directory will not be found.

This change modifies the non-root-dir-parsing loop of do_fat_read_at()
to detect if it's walked back to the root directory, and if so, jumps
back to the special case code that handles parsing of the root directory.

This change was tested using sandbox by executing:

./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/.."
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/../"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/../backup"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/../backup/"
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/../backup/.."
./u-boot -c "host bind 0 ../sd-p1.bin; ls host 0:0 /extlinux/../backup/../"
./u-boot -c "host bind 0 ../sd-p1.bin; load host 0:0 0 /bcm2835-rpi-cm.dtb"
./u-boot -c "host bind 0 ../sd-p1.bin; load host 0:0 0 /extlinux/../bcm2835-rpi-cm.dtb"
./u-boot -c "host bind 0 ../sd-p1.bin; load host 0:0 0 /backup/../bcm2835-rpi-cm.dtb"
./u-boot -c "host bind 0 ../sd-p1.bin; load host 0:0 0 /extlinux/..backup/../bcm2835-rpi-cm.dtb"
./u-boot -c "host bind 0 ../sd-p1.bin; load host 0:0 0 /extlinux/../backup/../bcm2835-rpi-cm.dtb"

(/extlinux and /backup are in different sectors so trigger some different
cases, and bcm2835-rpi-cm.dtb is in a sector of the root directory other
than the first).

In all honesty, this change is a bit of a hack, using goto and all.
However, as demonstrated above it appears to work well in practice, is
quite minimal, likely doesn't introduce any risk of regressions, and
hopefully doesn't introduce any maintenance issues.

The correct fix would be to collapse the root and non-root loops in
do_fat_read_at() and get_dentfromdir() into a single loop that has a
small special-case when moving from one sector to the next, to handle
the layout difference of root/non-root directories. AFAIK all other
aspects of directory parsing are identical. However, that's a much
larger change which needs significantly more thought before it's
implemented.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
4 years agogit-mailrc: add Alexey as ARC maintainer
Masahiro Yamada [Mon, 11 May 2015 10:44:01 +0000 (19:44 +0900)]
git-mailrc: add Alexey as ARC maintainer

It's easier to Cc him on ARC-releated patches.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
4 years agoMerge branch 'master' of http://git.denx.de/u-boot-sunxi
Tom Rini [Fri, 11 Sep 2015 17:05:05 +0000 (13:05 -0400)]
Merge branch 'master' of git.denx.de/u-boot-sunxi

4 years agoat91, taurus, smartweb: add dfu support
Heiko Schocher [Tue, 8 Sep 2015 09:52:52 +0000 (11:52 +0200)]
at91, taurus, smartweb: add dfu support

[root@pollux dfu-util]# ./src/dfu-util -l
dfu-util 0.8

Copyright 2005-2009 Weston Schmidt, Harald Welte and OpenMoko Inc.
Copyright 2010-2014 Tormod Volden and Stefan Schmidt
This program is Free Software and has ABSOLUTELY NO WARRANTY
Please report bugs to dfu-util@lists.gnumonks.org

Found DFU: [0908:02d2] ver=0212, devnum=119, cfg=1, intf=0, alt=0, name="Linux", serial="UNKNOWN"
[root@pollux dfu-util]#

Signed-off-by: Heiko Schocher <hs@denx.de>
Acked-by: Lukasz Majewski <l.majewski@samsung.com>
4 years agousb: gadget: at91_udc: add at91_udc into U-Boot
Heiko Schocher [Tue, 8 Sep 2015 09:52:51 +0000 (11:52 +0200)]
usb: gadget: at91_udc: add at91_udc into U-Boot

add U-Boot specific changes to the at91_udc linux driver,
so it works with U-Boot.

Signed-off-by: Heiko Schocher <hs@denx.de>
4 years agousb: gadget: at91_udc: port linux driver at91_udc
Heiko Schocher [Tue, 8 Sep 2015 09:52:50 +0000 (11:52 +0200)]
usb: gadget: at91_udc: port linux driver at91_udc

port at91_udc driver from linux:

original commit Message:
commit c94e289f195e0e13cf34d27f9338d28221a85751
Author: Arnd Bergmann <arnd@arndb.de>
Date:   Sat Apr 11 00:14:21 2015 +0200

    usb: gadget: remove incorrect __init/__exit annotations

    A recent change introduced a link error for the composite
    printer gadget driver:

    `printer_unbind' referenced in section `.ref.data' of drivers/built-in.o: defined in discarded section `.exit.text' of drivers/built-in.o

    Evidently the unbind function should not be marked __exit here,
    because it is called through a callback pointer that is not necessarily
    discarded, __composite_unbind() is indeed called from the error path of
    composite_bind(), which can never work for a built-in driver.

    Looking at the surrounding code, I found the same problem in all other
    composite gadget drivers in both the bind and unbind functions, as
    well as the udc platform driver 'remove' functions. Those will break
    if anyone uses the 'unbind' sysfs attribute to detach a device from a
    built-in driver.

    This patch removes the incorrect annotations from all the gadget
    drivers.

Signed-off-by: Heiko Schocher <hs@denx.de>
4 years agoARM: atmel: boards: use default CONFIG_SYS_PBSIZE
Bo Shen [Wed, 2 Sep 2015 10:00:06 +0000 (18:00 +0800)]
ARM: atmel: boards: use default CONFIG_SYS_PBSIZE

Entering the maximum number of characters defined by CONFIG_SYS_CBSIZE
into the console and hitting enter afterwards, causes a hang in the
system because CONFIG_SYS_PBSIZE is not capable of storing the extra
characters of the error message:
"Unknown command '' - try 'help'".

Use the default CONFIG_SYS_PBSIZE definition from config_fallbacks.h
to solve this problem.

Signed-off-by: Bo Shen <voice.shen@atmel.com>
Signed-off-by: Josh Wu <josh.wu@atmel.com>
4 years agoarm: at91: convert meesc board to generic board
Daniel Gorsulowski [Mon, 31 Aug 2015 05:32:25 +0000 (07:32 +0200)]
arm: at91: convert meesc board to generic board

Signed-off-by: Daniel Gorsulowski <daniel.gorsulowski@esd.eu>
Reviewed-by: Simon Glass <sjg@chromium.org>
[fix corrupt line wraps in patch]
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
4 years agotaurus: board updates
Heiko Schocher [Fri, 21 Aug 2015 16:55:07 +0000 (18:55 +0200)]
taurus: board updates

taurus changes:
- rename at91_spl_board_init to spl_board_init
  fixes problems with recovery button and nand erase sector 0
- adapt CONFIG_SPL_MAX_SIZE and CONFIG_SPL_BSS_MAX_SIZE
- add CONFIG_AT91_HW_WDT_TIMEOUT 15
- CONFIG_SF_DEFAULT_MODE SPI_MODE_3 not mode 0

Signed-off-by: Heiko Schocher <hs@denx.de>
4 years agoarm, at91: add axm extensions
Heiko Schocher [Fri, 21 Aug 2015 16:53:46 +0000 (18:53 +0200)]
arm, at91: add axm extensions

add extensions for the axm board:
- power on LED on power up
- press both recovery buttons on power up to enter
  recovery mode
- detect 64 MiB and 128 MiB ramsize
- PHY rest at reboot because of ATMEL bug
- use siemens update concept
- add axm default environment
- set CONFIG_SPL_MAX_SIZE to 15k

Signed-off-by: Heiko Schocher <hs@denx.de>
4 years agocorvus, dfu: add dfu support
Heiko Schocher [Fri, 21 Aug 2015 09:28:20 +0000 (11:28 +0200)]
corvus, dfu: add dfu support

add support for DFU on the corvus board.

Signed-off-by: Heiko Schocher <hs@denx.de>
4 years agoat91: corvus: board updates
Heiko Schocher [Fri, 21 Aug 2015 09:28:19 +0000 (11:28 +0200)]
at91: corvus: board updates

- rename at91_spl_board_init into spl_board_init
- use SZ_1X defines for sizes

Signed-off-by: Heiko Schocher <hs@denx.de>
4 years agosunxi_nand_spl: Add config parameter for 4KiB page sized NAND devices
Stefan Roese [Fri, 4 Sep 2015 12:43:00 +0000 (14:43 +0200)]
sunxi_nand_spl: Add config parameter for 4KiB page sized NAND devices

This patch adds support for NAND chips with 4KiB page size and 24/1024
ECC strength. Like the Micron MT29F32G08CBACAWP which is used on the
ICnova-A20 SoM.

Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
4 years agosunxi: Ensure that 'mksunxiboot' tool produces deterministic output
Siarhei Siamashka [Wed, 2 Sep 2015 23:36:39 +0000 (02:36 +0300)]
sunxi: Ensure that 'mksunxiboot' tool produces deterministic output

Currently some uninitialized padding bytes are written to the output
file, as can be confirmed with valgrind:

$ valgrind tools/mksunxiboot spl/u-boot-spl.bin spl/sunxi-spl.bin

==5581== Syscall param write(buf) points to uninitialised byte(s)
==5581==    at 0x4F0F940: __write_nocancel (in /lib64/libc-2.20.so)
==5581==    by 0x400839: main (in /tmp/u-boot/tools/mksunxiboot)
==5581==  Address 0xffeff5d3c is on thread 1's stack
==5581==  in frame #1, created by main (???)

This patch fixes the problem by clearing the whole structure instead
of just a portion of it.

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
4 years agosunxi: Fix wrong serial console setup in Forfun Q88DB tablet
Siarhei Siamashka [Tue, 8 Sep 2015 14:49:08 +0000 (17:49 +0300)]
sunxi: Fix wrong serial console setup in Forfun Q88DB tablet

The Forfun Q88DB tablet was unbootable since commit
b6006baf9c2553543e3384983d23d95efbf24fa6 ("sunxi: Move all
boards to the driver-model"). Appears that this is caused
by the wrong serial console setup in the SPL. The serial
console should use PG3/PG4 pins according to the FEX file.

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
4 years agosun5i: Add A10s-Wobo-i5 defconfig and dts
Jelle van der Waa [Mon, 7 Sep 2015 20:43:57 +0000 (22:43 +0200)]
sun5i: Add A10s-Wobo-i5 defconfig and dts

The Wobo i5 top set box is a somewhat curious A10s based top set box,
it uses an AXP209 rather then the AXP152 usually used in combination
with the A10s. It has an ethernet phy connected to PORTD rather then
PORTA, and its built-in usb wifi is connected via the otg controller.

The dts file changes are identical to the changes submitted to the
upstream kernel.

Signed-off-by: Jelle van der Waa <jelle@vdwaa.nl>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
4 years agosun5i: Add q8_a13_tablet defconfig and dts
Hans de Goede [Sat, 5 Sep 2015 15:01:31 +0000 (17:01 +0200)]
sun5i: Add q8_a13_tablet defconfig and dts

This commits adds a generic support for q8 formfactor a13 based tablets.

These tablets ship in many variants, with the difference mainly being the
touchscreen controller / accelerometer / wifi chip used.

The wifi is USB based, and thus not listed in devicetree.

ATM the kernel does not support the touchscreen / accelerometer on these
devices. In the future we may need multiple configs with different
CONFIG_DEFAULT_DEVICE_TREE settings, this depends on how we solve the
hw differences on the kernel side.

For now this will suffice.

The dts files are identical to the dts files submitted to the upstream
kernel for these tablets.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
4 years agosun4i: Add dts and defconfig for iNet-1 based tablets
Hans de Goede [Wed, 9 Sep 2015 07:46:30 +0000 (09:46 +0200)]
sun4i: Add dts and defconfig for iNet-1 based tablets

The iNet-tek iNet-1 PCB is a PCB found in various generic 10.1" 1024x600
A10 based tablets such as the Point of View Protab2 XXL and the
Cherry M1007.

This patch has been tested on both rev2 and rev5 of this board / these
tablets.

These tablets feature the usual connectors: headphone, mini hdmi,
power-barrel, mini-usb and a micro-sd slot.

The dts is identical to the dts submitted to the upstream kernel.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
4 years agosun4i: Add defconfig and dts for inet9f-rev03 based tablets
Hans de Goede [Tue, 8 Sep 2015 20:28:30 +0000 (22:28 +0200)]
sun4i: Add defconfig and dts for inet9f-rev03 based tablets

The inet9f-rev03 pcb is specially designed for gaming tablets, such as
the qware tb-g100 tablet.

These 7" tablets feature a dpad, firebuttons and 2 joysticks on the sides
of the screen.

Besides this they have the usual connectors: power-barrel, mini usb,
mini hdmi, headphone and micro-sd slot.

The dts is identical to the dts submitted to the upstream kernel.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
4 years agosun4i: Add defconfig and dts for the pov protab2-ips9 tablet
Hans de Goede [Sat, 5 Sep 2015 08:50:56 +0000 (10:50 +0200)]
sun4i: Add defconfig and dts for the pov protab2-ips9 tablet

The Point of View protab2-ips9 is a tablet with a 9" ips 1024x768 lcd
screen, microsd slot, headphones, mini hdmi, mini usb b and power barrel
connectors.

The dts file is identical to the one submitted to the upstream kernel.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
4 years agonet: designware: Fix build warnings
Bin Meng [Thu, 3 Sep 2015 12:37:29 +0000 (05:37 -0700)]
net: designware: Fix build warnings

When building dm version of designware eth driver on a platform
with 64-bit phys_addr_t, it reports the following warnings:

  drivers/net/designware.c: In function 'designware_eth_probe':
  drivers/net/designware.c:599:2:
    warning: format '%lx' expects argument of type 'long unsigned int',
    but argument 3 has type 'phys_addr_t' [-Wformat]
  drivers/net/designware.c:600:21:
    warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
  drivers/net/designware.c:601:21:
    warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]

This commit fixes the build warnings.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
4 years agox86: galileo: Convert to use CONFIG_DM_USB
Bin Meng [Thu, 3 Sep 2015 12:37:28 +0000 (05:37 -0700)]
x86: galileo: Convert to use CONFIG_DM_USB

Move to driver model for USB on Intel Galileo.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: quark: Add USB PHY initialization support
Bin Meng [Thu, 3 Sep 2015 12:37:27 +0000 (05:37 -0700)]
x86: quark: Add USB PHY initialization support

USB PHY needs to be properly initialized per Quark firmware writer
guide, otherwise the EHCI controller on Quark SoC won't work.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: Convert to use driver model pci on quark/galileo
Bin Meng [Thu, 3 Sep 2015 12:37:26 +0000 (05:37 -0700)]
x86: Convert to use driver model pci on quark/galileo

Move to driver model pci for Intel quark/galileo.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: Enable PCIe controller on quark/galileo
Bin Meng [Thu, 3 Sep 2015 12:37:25 +0000 (05:37 -0700)]
x86: Enable PCIe controller on quark/galileo

Quark SoC holds the PCIe controller in reset following a power on.
U-Boot needs to release the PCIe controller from reset. The PCIe
controller (D23:F0/F1) will not be visible in PCI configuration
space and any access to its PCI configuration registers will cause
system hang while it is held in reset.

Enable PCIe controller per Quark firmware writer guide.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: quark: Avoid chicken and egg problem
Bin Meng [Thu, 3 Sep 2015 12:37:24 +0000 (05:37 -0700)]
x86: quark: Avoid chicken and egg problem

If we convert to use driver model pci on quark, we will encounter
some chicken and egg problems like below:

- To enable PCIe root ports, we need program some registers on the
  message bus via pci bus. With driver model, the first time to
  access pci bus, the pci enumeration process will be triggered.
  But without first enabling PCIe root ports, pci enumeration
  just hangs when scanning PCIe root ports.
- Similar situation happens when trying to access GPIO from the
  PCIe enabling codes, as GPIO requires its block base address
  to be assigned via a pci configuration register in the bridge.

To avoid such dilemma, replace all pci calls in the quark codes
to use the local version which does not go through driver model.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: quark: Optimize MRC execution time
Bin Meng [Thu, 3 Sep 2015 12:37:23 +0000 (05:37 -0700)]
x86: quark: Optimize MRC execution time

Intel Quark SoC has a low end x86 processor with only 400MHz
frequency. Currently it takes about 15 seconds for U-Boot to
boot to shell and the most time consuming part is with MRC,
which is about 12 seconds. MRC programs lots of registers on
the SoC internal message bus indirectly accessed via pci bus.

To speed up the boot, create an optimized version of pci config
read/write dword routines which directly operate on PCI I/O ports.
These two routines are inlined to provide better performance too.
Now it only takes about 3 seconds to finish MRC, which is really
fast (4 times faster than before).

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agodm: pci: Allow a PCI bus to be found without an alias
Simon Glass [Tue, 1 Sep 2015 00:55:35 +0000 (18:55 -0600)]
dm: pci: Allow a PCI bus to be found without an alias

At present, until a PCI bus is probed, it cannot be found by its sequence
number unless it has an alias. This is the same with any device.

However with PCI this is more annoying than usual, since bus 0 is always the
same device.

Add a function that tries a little harder to locate PCI bus 0. This means
that PCI enumeration will happen automatically on the first access.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
4 years agobuildman: Improve the config comparison feature
Simon Glass [Wed, 26 Aug 2015 03:52:14 +0000 (21:52 -0600)]
buildman: Improve the config comparison feature

At present buildman can compare configurations between commits but the
feature is less useful than it could be. There is no summary by architecture
and changes are not reported on a per-board basis.

Correct these deficiencies so that it is possible to see exactly what is
changing for any number of boards.

Note that 'buildman -b <branch> -C' is recommended for any build where you
will be comparing configuration. Without -C the correct configuration will
not be reported since changes will often not be picked up.

Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agoRevert "patman: use -D option for git format-patch"
Masahiro Yamada [Sun, 30 Aug 2015 16:23:32 +0000 (01:23 +0900)]
Revert "patman: use -D option for git format-patch"

This reverts commit 19b4a3369876f9215e2b861f211e8df1a75e26ca.

Since that commit, patman generates useless patches for file removal;
"git format -D" prints only the header but not the diff when deleting
files, and "git am" always refuses such patches.

The following is the quotation from "man git-format-patch":

  -D, --irreversible-delete
    Omit the preimage for deletes, i.e. print only the header but
    not the diff between the preimage and /dev/null. The resulting
    patch is not meant to be applied with patch nor git apply; this
    is solely for people who want to just concentrate on reviewing
    the text after the change. In addition, the output obviously
    lack enough information to apply such a patch in reverse, even
    manually, hence the name of the option.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: bayleybay: Convert to use more dm drivers
Bin Meng [Fri, 28 Aug 2015 09:22:39 +0000 (02:22 -0700)]
x86: bayleybay: Convert to use more dm drivers

Move to driver model for USB and ETH on Intel Bayley Bay.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: coreboot: Convert to use more dm drivers
Bin Meng [Fri, 28 Aug 2015 09:22:38 +0000 (02:22 -0700)]
x86: coreboot: Convert to use more dm drivers

Move to driver model for RTC, USB and ETH on coreboot.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: crownbay: Enable CONFIG_PCH_GBE
Bin Meng [Fri, 28 Aug 2015 05:25:59 +0000 (22:25 -0700)]
x86: crownbay: Enable CONFIG_PCH_GBE

Now that we have converted the pch_gbe driver to driver moel,
enable it on Intel Crown Bay board.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agonet: pch_gbe: Add Kconfig option
Bin Meng [Fri, 28 Aug 2015 05:25:58 +0000 (22:25 -0700)]
net: pch_gbe: Add Kconfig option

Add Kconfig option in preparation for moving board to use Kconfig.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
4 years agonet: pch_gbe: Convert to driver model
Bin Meng [Fri, 28 Aug 2015 05:25:57 +0000 (22:25 -0700)]
net: pch_gbe: Convert to driver model

This commit converts pch_gbe ethernet driver to driver model.

Since this driver is only used by Intel Crown Bay board, the
conversion does not keep the non-dm version.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agox86: crownbay: Convert to use CONFIG_DM_ETH for E1000
Bin Meng [Fri, 28 Aug 2015 05:25:56 +0000 (22:25 -0700)]
x86: crownbay: Convert to use CONFIG_DM_ETH for E1000

Since E1000 driver has been converted to driver model, enable it
on Intel Crown Bay. But the Intel Topcliff GbE driver has not been
converted to driver model yet, disable it for now.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agox86: crownbay: Convert to use CONFIG_DM_USB
Bin Meng [Fri, 28 Aug 2015 05:25:55 +0000 (22:25 -0700)]
x86: crownbay: Convert to use CONFIG_DM_USB

Move to driver model for USB on Intel Crown Bay.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
4 years agodm: eth: Correctly detect alias in eth_get_dev_by_name()
Bin Meng [Fri, 28 Aug 2015 05:25:54 +0000 (22:25 -0700)]
dm: eth: Correctly detect alias in eth_get_dev_by_name()

When given a device name string, we should test to see if it is
really an alias like "eth#".

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodm: test: Add a new test case for dm_test_eth_rotate
Bin Meng [Fri, 28 Aug 2015 05:25:53 +0000 (22:25 -0700)]
dm: test: Add a new test case for dm_test_eth_rotate

Add one more ethernet device node in the sandbox test device tree,
with name 'sbe5'. This is to support a new test case for testing
network device rotation.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodm: eth: Do not print misleading "Net Initialization Skipped"
Bin Meng [Fri, 28 Aug 2015 05:25:52 +0000 (22:25 -0700)]
dm: eth: Do not print misleading "Net Initialization Skipped"

With driver model, board_eth_init() or cpu_eth_init() is not a must.
Thus we don't need print a misleading "Net Initialization Skipped".

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agonet: Revert "tftp: adjust settings to be suitable for 100Mbit ethernet"
Bin Meng [Fri, 28 Aug 2015 05:25:51 +0000 (22:25 -0700)]
net: Revert "tftp: adjust settings to be suitable for 100Mbit ethernet"

Commit 620776d "tftp: adjust settings to be suitable for 100Mbit ethernet"
causes the following error message when trying to load a file using 'tftp'
command via a tftp server.

    TFTP error: 'Unsupported option(s) requested' (8)

This is due to with commit 620776d changes, the tftp option 'timeout'
value is now set to zero which is an invalid value as per RFC2349 [1].
Valid values range between "1" and "255" seconds, inclusive. With some
tftp servers that strictly implement the RFC requirement, it reports
such an error message.

Revert commit 620776d for RFC compliance.

[1] https://www.ietf.org/rfc/rfc2349.txt

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agox86: panther: Add PCI and video configuration
Simon Glass [Fri, 28 Aug 2015 01:54:48 +0000 (19:54 -0600)]
x86: panther: Add PCI and video configuration

Add a PCI node to the device tree. This allows SPI flash and SATA to work
correctly. Also configure the video to come up correctly even though there
is no keyboard.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
4 years agox86: bayleybay: Change default vga bios rom address
Bin Meng [Thu, 27 Aug 2015 15:38:17 +0000 (08:38 -0700)]
x86: bayleybay: Change default vga bios rom address

With multiple microcode blobs included, the generated u-boot-dtb.bin
leaves merely several hundred bytes before the vga bios. Change the
vga bios rom address to make a bigger room for u-boot-dtb.bin.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
4 years agox86: doc: Change to use CONFIG_VGA_BIOS_ADDR
Bin Meng [Thu, 27 Aug 2015 15:38:16 +0000 (08:38 -0700)]
x86: doc: Change to use CONFIG_VGA_BIOS_ADDR

CONFIG_X86_OPTION_ROM_ADDR has been renamed to CONFIG_VGA_BIOS_ADDR.
Update the doc to refer to the new name.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
4 years agox86: ifdtool: Support checking region overlap before U-Boot
Bin Meng [Thu, 27 Aug 2015 15:38:15 +0000 (08:38 -0700)]
x86: ifdtool: Support checking region overlap before U-Boot

We have the capability to check regions written after U-Boot that
do not overlap. Since regions can also be written before U-Boot,
add such check for these too.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Andy Pont <andy.pont@sdcsystems.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
4 years agoPrepare v2015.10-rc3
Tom Rini [Mon, 7 Sep 2015 12:56:35 +0000 (08:56 -0400)]
Prepare v2015.10-rc3

Signed-off-by: Tom Rini <trini@konsulko.com>
4 years agoMerge git://git.denx.de/u-boot-usb
Tom Rini [Mon, 7 Sep 2015 12:56:23 +0000 (08:56 -0400)]
Merge git://git.denx.de/u-boot-usb

4 years agoMerge git://git.denx.de/u-boot-socfpga
Tom Rini [Mon, 7 Sep 2015 12:56:08 +0000 (08:56 -0400)]
Merge git://git.denx.de/u-boot-socfpga

4 years agousb: ehci: remember init mode
Stephen Warren [Thu, 20 Aug 2015 23:38:05 +0000 (17:38 -0600)]
usb: ehci: remember init mode

When an EHCI device is registered in device mode, the HW isn't actually
initialized at all, and hence isn't left in a running state. Consequently,
when the device is deregistered, ehci_shutdown() will fail, since the HW
bits it expects to see set in response to its shutdown requests will not
be sent, and the message "EHCI failed to shut down host controller." will
be printed.

Fix ehci-hcd.c to remember whether the device was registered in host or
device mode, and only call ehci_shutdown() for host mode registrations.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
4 years agodfu: tftp: Kconfig: Enable DFU_TFTP support on the am335x_boneblack_defconfig
Lukasz Majewski [Sun, 23 Aug 2015 22:21:50 +0000 (00:21 +0200)]
dfu: tftp: Kconfig: Enable DFU_TFTP support on the am335x_boneblack_defconfig

This commit enables support for DFU_TFTP on the am335x bone black device.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodfu: tftp: Kconfig: Add Kconfig entry for dfu tftp feature
Lukasz Majewski [Sun, 23 Aug 2015 22:21:49 +0000 (00:21 +0200)]
dfu: tftp: Kconfig: Add Kconfig entry for dfu tftp feature

The dfu tftp feature can be now enabled via Kconfig. This
commit provides necessary code for it.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodfu: command: Extend "dfu" command to handle receiving data via TFTP
Lukasz Majewski [Sun, 23 Aug 2015 22:21:48 +0000 (00:21 +0200)]
dfu: command: Extend "dfu" command to handle receiving data via TFTP

The "dfu" command has been extended to support transfers via TFTP protocol.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agoupdate: tftp: dfu: Extend update_tftp() function to support DFU
Lukasz Majewski [Sun, 23 Aug 2015 22:21:47 +0000 (00:21 +0200)]
update: tftp: dfu: Extend update_tftp() function to support DFU

This code allows using DFU defined mediums for storing data received via
TFTP protocol.

It reuses and preserves functionality of legacy code at common/update.c.

The update_tftp() function now accepts parameters - namely medium device
name and its number (e.g. mmc 1).

Without this information passed old behavior is preserved.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodfu: tftp: update: Add dfu_write_from_mem_addr() function
Lukasz Majewski [Sun, 23 Aug 2015 22:21:46 +0000 (00:21 +0200)]
dfu: tftp: update: Add dfu_write_from_mem_addr() function

This function allows writing via DFU data stored from fixed buffer address
(like e.g. loadaddr env variable).

Such predefined buffers are used in the update_tftp() code. In fact this
function is a wrapper on the dfu_write() and dfu_flush().

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodfu: tftp: update: Provide tftp support for the DFU subsystem
Lukasz Majewski [Sun, 23 Aug 2015 22:21:45 +0000 (00:21 +0200)]
dfu: tftp: update: Provide tftp support for the DFU subsystem

This commit adds initial support for using tftp for downloading and
upgrading firmware on the device.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agotftp: update: Allow some parts of the code to be reused when CONFIG_SYS_NO_FLASH...
Lukasz Majewski [Sun, 23 Aug 2015 22:21:44 +0000 (00:21 +0200)]
tftp: update: Allow some parts of the code to be reused when CONFIG_SYS_NO_FLASH is set

Up till now it was impossible to use code from update.c when system
was not equipped with raw FLASH memory.
Such behavior prevented DFU from reusing this code.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
4 years agonet: tftp: Move tftp.h file from ./net to ./include/net
Lukasz Majewski [Sun, 23 Aug 2015 22:21:43 +0000 (00:21 +0200)]
net: tftp: Move tftp.h file from ./net to ./include/net

This change gives the ability to reuse the <tftp.h> header file by other
subsystems (like e.g. dfu).

Without this change compilation error emerges for the legacy update.c file.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agodoc: dfu: tftp: README entry for TFTP extension of DFU
Lukasz Majewski [Sun, 23 Aug 2015 22:21:42 +0000 (00:21 +0200)]
doc: dfu: tftp: README entry for TFTP extension of DFU

Documentation file for DFU extension. With this functionality it is now
possible to transfer FIT images with firmware updates via TFTP and use
DFU backend for storing them.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
4 years agoh2200: Fix build error
Kishon Vijay Abraham I [Fri, 21 Aug 2015 05:32:04 +0000 (11:02 +0530)]
h2200: Fix build error

Commit <8bfc288c3955> ("usb: gadget: ether: Perform board
initialization from ethernet gadget driver") added board_usb_init
and board_usb_cleanup in ethernet gadget driver. But h2200 board
didn't have board_usb_init and board_usb_cleanup implementations.

This introduced the following build errors
+drivers/usb/gadget/built-in.o: In function `usb_eth_halt':
+drivers/usb/gadget/ether.c:2498: undefined reference to `board_usb_cleanup'
+drivers/usb/gadget/built-in.o: In function `usb_eth_init':
+drivers/usb/gadget/ether.c:2316: undefined reference to `board_usb_init'

Fix it here by adding empty board_usb_init and board_usb_cleanup
functions in h2200.c.

Fixes: <8bfc288c3955> ("usb: gadget: ether: Perform board
initialization from ethernet gadget driver")

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
4 years agodfu: Delete superfluous initialization of the dfu_buf_size static variable
Lukasz Majewski [Wed, 8 Jul 2015 21:43:18 +0000 (23:43 +0200)]
dfu: Delete superfluous initialization of the dfu_buf_size static variable

After extension of the dfu_get_buf() to also setup (implicitly) the dfu_buf_size
variable it is not needed to set dfu_buf_size to CONFIG_SYS_DFU_DATA_BUF_SIZE.

This variable is set in the dfu_get_buf() by not only considering
CONFIG_SYS_DFU_DATA_BUF but more importantly the "dfu_bufsiz" env variable.
Therefore, dfu_get_buf() should be used for initialization.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Reviewed-by: Przemyslaw Marczak <p.marczak@samsung.com>
4 years agodfu:tests: Modify dfu_gadget_test.sh to accept USB device vendor:product ID
Lukasz Majewski [Sat, 15 Aug 2015 09:23:45 +0000 (11:23 +0200)]
dfu:tests: Modify dfu_gadget_test.sh to accept USB device vendor:product ID

dfu-util allows filtering on USB device vendor:product ID by using
the -d flag (-d 0451:d022).
Such option is very handy when many DFU devices are connected to a single
host PC. This commit allows testing when above situation emerges.

Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Lukasz Majewski <l.majewski@majess.pl>
Test HW - AM335x Beagle Bone Black
NOTE: Max size of file to transfer: 2MiB

4 years agousb: gadget: f_thor: Allocate request up to THOR_PACKET_SIZE
Siva Durga Prasad Paladugu [Wed, 15 Apr 2015 11:42:19 +0000 (13:42 +0200)]
usb: gadget: f_thor: Allocate request up to THOR_PACKET_SIZE

Allocate request up to THOR_PACKET_SIZE not the ep0->maxpacket
as the descriptors data depend on the number of descriptors
and this 64 bytes were not enough and the buffer might overflow
which results in memalign failures later.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
4 years agof_thor: Dont perform reset at the end of thor
Siva Durga Prasad Paladugu [Thu, 16 Apr 2015 08:38:34 +0000 (10:38 +0200)]
f_thor: Dont perform reset at the end of thor

Dont perform reset at the end of thor download
if configured to do reset off.
Reset may not be required in all cases and hence
provided an option to do so.

The case would be to download the images to DDR instead
of flash device.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
4 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>
4 years agoarc: make AXS101 default platform
Alexey Brodkin [Mon, 24 Aug 2015 15:49:37 +0000 (18:49 +0300)]
arc: make AXS101 default platform

This fixes building in automated flow that doesn't use defconfigs.

See discussion on that topic here:
 http://patchwork.ozlabs.org/patch/502558/

See similar patches for other architectures/platforms  here:
 [1] http://git.denx.de/?p=u-boot.git;a=commit;h=ff560a13056a565a4e9ce1761bd04276a3cace88
 [2] http://git.denx.de/?p=u-boot.git;a=commit;h=589907e2c187ec69b351c38ccda36730d25ab5d6

And while at it add missing shell prompt to axs103.

Cc: Tom Rini <trini@konsulko.com>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
4 years agonet: altera_tse: Zap unused variable
Marek Vasut [Fri, 4 Sep 2015 09:50:28 +0000 (11:50 +0200)]
net: altera_tse: Zap unused variable

Zap variable which is set but never used.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Tested-by: Thomas Chou <thomas@wytron.com.tw>
4 years agoarm: socfpga: Add support for the Terasic DE-0 Atlas board
Dinh Nguyen [Tue, 1 Sep 2015 22:41:52 +0000 (17:41 -0500)]
arm: socfpga: Add support for the Terasic DE-0 Atlas board

Add support for the Terasic DE0-Nano/Atlas-SoC Kit, which is a CycloneV
based board. The board can boot from SD/MMC. Ethernet is also supported.

Signed-off-by: Dinh Nguyen <dinguyen@opensource.altera.com>
4 years agoarm: socfpga: Add support for DENX MCV SoM and MCVEVK board
Marek Vasut [Sun, 2 Aug 2015 23:37:28 +0000 (01:37 +0200)]
arm: socfpga: Add support for DENX MCV SoM and MCVEVK board

Add support for DENX MCV SoM, which is CycloneV based and the
associated DENX MCVEVK baseboard. The board can boot from eMMC.
Ethernet and USB is supported.

Signed-off-by: Marek Vasut <marex@denx.de>
4 years agoarm: socfpga: Add support for Terasic SoCkit board
Marek Vasut [Sun, 21 Jun 2015 15:28:53 +0000 (17:28 +0200)]
arm: socfpga: Add support for Terasic SoCkit board

Add support for Terasic SoCkit, which is CycloneV based board.
The board can boot either from SD/MMC or QSPI. Ethernet is also
supported.

Signed-off-by: Marek Vasut <marex@denx.de>
4 years agoarm: socfpga: Do not call board_init_r() from board_init_f()
Marek Vasut [Sat, 29 Aug 2015 08:16:20 +0000 (10:16 +0200)]
arm: socfpga: Do not call board_init_r() from board_init_f()

Instead of calling board_init_r() directly from board_init_f(), just
return from board_init_f(). This will make the code continue executing
in crt0.S _main(), from which the board_init_r() is called. This patch
aligns the SoCFPGA SPL with the correct SPL design as well as reduces
the stack utilisation slightly.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
4 years agoarm: socfpga: Zap OF_CONTROL checks, it's always enabled
Marek Vasut [Wed, 19 Aug 2015 21:23:53 +0000 (23:23 +0200)]
arm: socfpga: Zap OF_CONTROL checks, it's always enabled

The CONFIG_OF_CONTROL and CONFIG_SPL_OF_CONTROL is always enabled
on Altera SoCFPGA, remove the unnecessary checks.

Signed-off-by: Marek Vasut <marex@denx.de>
4 years agoarm: socfpga: Always enable OF_CONTROL and SPL_OF_CONTROL
Marek Vasut [Wed, 19 Aug 2015 21:23:52 +0000 (23:23 +0200)]
arm: socfpga: Always enable OF_CONTROL and SPL_OF_CONTROL

The SoCFPGA probes mostly from OF and the OF is mandatory both in
U-Boot itself and U-Boot SPL. Enable it by default.

Signed-off-by: Marek Vasut <marex@denx.de>
4 years agoarm: socfpga: Assure ISWGRP 0 and 1 are inited
Marek Vasut [Mon, 24 Aug 2015 09:51:46 +0000 (11:51 +0200)]
arm: socfpga: Assure ISWGRP 0 and 1 are inited

This fix makes sure that the ISWGRP0 and ISWGRP1 registers are
correctly inited. In case those registers are not initialized,
it is not possible to access the registers synthesised in the
FPGA through the bridges. Any such access produces data abort.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
4 years agommc: dw_mmc: Probe the MMC from OF
Marek Vasut [Sat, 25 Jul 2015 08:48:14 +0000 (10:48 +0200)]
mmc: dw_mmc: Probe the MMC from OF

Rework the driver to probe the MMC controller from Device Tree
and make it mandatory. There is no longer support for probing
from the ancient qts-generated header files.

This patch now also removes previous temporary workaround.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@konsulko.com>
4 years agodoc: document the fdtdir PXE command
Stefan Brüns [Sun, 30 Aug 2015 17:10:58 +0000 (19:10 +0200)]
doc: document the fdtdir PXE command

Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
4 years agodoc: Fix inconsistent filename in PXE config example
Stefan Brüns [Sun, 30 Aug 2015 17:10:59 +0000 (19:10 +0200)]
doc: Fix inconsistent filename in PXE config example

The default config includes base.menu, not linux.list

Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
4 years agoMerge git://git.denx.de/u-boot-rockchip
Tom Rini [Thu, 3 Sep 2015 18:57:09 +0000 (14:57 -0400)]
Merge git://git.denx.de/u-boot-rockchip

4 years agorockchip: Put README image creation commands on one line
Simon Glass [Sun, 30 Aug 2015 22:55:52 +0000 (16:55 -0600)]
rockchip: Put README image creation commands on one line

It is easier to paste these into the command line if they are a single
common. Use line continuation instead of separate lines.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Update todo in README.rockchip
Sjoerd Simons [Sun, 30 Aug 2015 22:55:51 +0000 (16:55 -0600)]
rockchip: Update todo in README.rockchip

MMC support works now, so it can be dropped from the todo

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Drop first 32kb of zeros from the rkSD image type
Sjoerd Simons [Sun, 30 Aug 2015 22:55:50 +0000 (16:55 -0600)]
rockchip: Drop first 32kb of zeros from the rkSD image type

Instead of creating a rockchip SPL SD card image with 32KB of zeros
which can be written to the start of an SD card, create the images with
only the useful data that should be written to an offset of 32KB on the
SD card.

The first 32 kilobytes aren't needed for bootup and only serve as
convenient way of accidentally obliterating your partition table.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agoarm: Turn of d-cache before i-cache
Sjoerd Simons [Sun, 30 Aug 2015 22:55:49 +0000 (16:55 -0600)]
arm: Turn of d-cache before i-cache

Booting the kernel fails on RK3288 (and probably other rockchip SoCs)
when the i-cache is disabled/flushed before d-cache.

I have not investigated whether this is due to U-Boot hanging or whether
it's very early in the linux boot, but following the approach of the
various rockchip U-Boot forks (first disable d-cache then i-cache) makes
things work.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add config_distro_bootcmd support
Sjoerd Simons [Sun, 30 Aug 2015 22:55:48 +0000 (16:55 -0600)]
rockchip: Add config_distro_bootcmd support

Now that MMC works in U-Boot add config distro command support to start
Linux in a standard fashion. One oddity here is that linux fails to load
when the fdt is relocated to above 512MB, so set fdt_high to make sure it's
loaded below that.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Turn off CONFIG_SPL_LED for firefly
Sjoerd Simons [Sun, 30 Aug 2015 22:55:47 +0000 (16:55 -0600)]
rockchip: Turn off CONFIG_SPL_LED for firefly

With LED support enabled the SPL easily goes over the size limit (e.g.
with both Debians gcc 4.9 and 5.2 cross-compilers). Turn off LED support
in the SPL to reduce the size just enough for those compilers.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Tweaked commit subject to remove _SUPPORT
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Disable sdio mmc slot on rk3288-firefly
Sjoerd Simons [Sun, 30 Aug 2015 22:55:46 +0000 (16:55 -0600)]
rockchip: Disable sdio mmc slot on rk3288-firefly

U-Boot can't use the sdio card so turn it of to prevent things getting
confused/struck when trying to use the card as storage.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agommc: Probe DM based mmc devices in u-boot
Sjoerd Simons [Sun, 30 Aug 2015 22:55:45 +0000 (16:55 -0600)]
mmc: Probe DM based mmc devices in u-boot

During mmc initialize probe all devices with the MMC Uclass if build
with CONFIG_DM_MMC

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agodoc: Fix reference to Rock pro when Rock 2 is meant
Sjoerd Simons [Sun, 30 Aug 2015 22:55:44 +0000 (16:55 -0600)]
doc: Fix reference to Rock pro when Rock 2 is meant

The Radxa Rock pro board is rk3188 based and thus won't work with U-Boot
built for RK3288. Change the documentation to refer to the intended
board, the Radxa Rock 2, which is an RK3288-based design very similar to
the firefly

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add a simple README
Simon Glass [Sun, 30 Aug 2015 22:55:43 +0000 (16:55 -0600)]
rockchip: Add a simple README

Add a few notes on how to try out the Rockchip support so far.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add basic support for jerry
Simon Glass [Sun, 30 Aug 2015 22:55:42 +0000 (16:55 -0600)]
rockchip: Add basic support for jerry

This builds and displays an SPL message, but does not function beyond that.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add basic support for firefly-rk3288
Simon Glass [Sun, 30 Aug 2015 22:55:41 +0000 (16:55 -0600)]
rockchip: Add basic support for firefly-rk3288

The Firefly RK3288 is a suitable target board for initial mainline Rockchip
support. It includes a good set of peripherals, a recent SoC and it is
readily available.

This adds only some basic files required to allow the baord to display a
serial message in SPL and hang.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add SPI driver
Simon Glass [Wed, 2 Sep 2015 01:19:37 +0000 (19:19 -0600)]
rockchip: Add SPI driver

Add a SPI driver for the Rockchip RK3288, using driver model. It should work
for other Rockchip SoCs also.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add I2C driver
Simon Glass [Sun, 30 Aug 2015 22:55:39 +0000 (16:55 -0600)]
rockchip: Add I2C driver

Add an I2C driver for the Rockchip RK3288, using driver model. It should work
for other Rockchip SoCs also.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 years agorockchip: Add core SoC start-up code
Simon Glass [Sun, 30 Aug 2015 22:55:38 +0000 (16:55 -0600)]
rockchip: Add core SoC start-up code

Add code for starting up U-Boot SPL and U-Boot proper. This is generic and
makes use of devices provided by the board- or SoC-specific code.

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