]> git.sur5r.net Git - u-boot/log
u-boot
9 years agoTegra: clocks: Add 38.4MHz OSC support for T210 use
Tom Warren [Mon, 22 Jun 2015 20:03:44 +0000 (13:03 -0700)]
Tegra: clocks: Add 38.4MHz OSC support for T210 use

Added 38.4MHz/48MHz entries to pll_x_table for CPU PLL. Needs
to be measured - should be close to 700MHz (1.4G/2).

Note that some freqs aren't in the PLLU table in T210 TRM
(13, 26MHz), so I used the 12MHz table entry for them. They
shouldn't be selected since they're not viable T210 OSC freqs.

Since there are now 2 new OSC defines, all tables (pll_x_table,
PLLU) had to increase by two entries, but since 38.4/48MHz are
not viable osc freqs on T20/30/114, etc, they're just set to 0.

Signed-off-by: Tom Warren <twarren@nvidia.com>
9 years agoT210: P2571: Turn CPU fan on
Tom Warren [Fri, 20 Feb 2015 19:22:22 +0000 (12:22 -0700)]
T210: P2571: Turn CPU fan on

CPU board (E2530) has a fan - turn it on via GPIO to keep
the SoC cool.

Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
9 years agoT210: P2571: Restore USB gadget mode (ums)
Tom Warren [Tue, 28 Jul 2015 22:16:38 +0000 (15:16 -0700)]
T210: P2571: Restore USB gadget mode (ums)

The tegra-common-usb-gadget.h include was causing
some build problems in ci_udc.c with a 64-bit gcc
in an earlier version of the T210 patches, but it
is working fine now, so restore it.

Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
9 years agoT210: P2571: Enable SD-card power via PMIC LDO2
Tom Warren [Tue, 28 Jul 2015 22:13:17 +0000 (15:13 -0700)]
T210: P2571: Enable SD-card power via PMIC LDO2

This was done in the 32-bit AVP loader (SPL) but is
board-specific so should be moved to the CPU portion.

Signed-off-by: Tom Warren <twarren@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
9 years agoconfigs: Remove CONFIG_SERIAL_MULTI
Stephen Warren [Wed, 5 Aug 2015 17:51:38 +0000 (11:51 -0600)]
configs: Remove CONFIG_SERIAL_MULTI

This config option isn't used anywhere at all. Remove all places that
define/enable the option.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
9 years agoMerge git://git.denx.de/u-boot-x86
Tom Rini [Wed, 5 Aug 2015 18:12:37 +0000 (14:12 -0400)]
Merge git://git.denx.de/u-boot-x86

9 years agox86: Document how to write PIRQ information in the device tree
Bin Meng [Mon, 3 Aug 2015 03:33:35 +0000 (20:33 -0700)]
x86: Document how to write PIRQ information in the device tree

Document the development flow on figuring out PIRQ information
during the U-Boot porting.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agopci: Remove DEBUG from pci_compat.c
Bin Meng [Mon, 3 Aug 2015 03:33:34 +0000 (20:33 -0700)]
pci: Remove DEBUG from pci_compat.c

Remove DEBUG in drivers/pci/pci_compat.c.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agousb: Fix device detection code
Marek Vasut [Wed, 5 Aug 2015 01:19:22 +0000 (03:19 +0200)]
usb: Fix device detection code

The code in question polls an USB port status via USB_REQ_GET_STATUS
to determine whether there is a device on the port or not. The way to
figure that out is to check two bits. Those are wPortChange[0] and
wPortStatus[0].

The wPortChange[0] indicates whether some kind of a connection status
change happened on a port (a device was plugged or unplugged). The
wPortStatus[0] bit indicates the status of the connection (plugged or
unplugged).

The current code tests whether wPortChange[0] == wPortStatus[0] and
if that's the case, considers the loop polling for the presence of a
USB device on port finished.

This works for most USB sticks, since they come up really quickly and
trigger the USB port change detection before the first iteration of the
detection loop happens. Thus, both wPortChange[0] and wPortStatus[0]
are set to 1 and thus equal. The loop is existed in it's first iteration
and the stick is detected correctly.

The problem is with some obscure USB sticks, which take some time before
they pop up on the bus after the port was enabled. In this case, both
the wPortChange[0] and wPortStatus[0] are 0. They are equal again, so
the loop again exits in the first iteration, but this is incorrect, as
such USB stick didn't have the opportunity to get detected on the bus.

Rework the code such, that it checks for wPortChange[0] first to test
if any connection change happened at all. If no change occured, keep
polling. If a change did occur, test the wPortStatus[0] to see there is
some device present on the port and only if this is the case, break out
of the polling loop.

This patch also trims down the duration of the polling loop from 10s
per port to 1s per port. This is still annoyingly long, but there is
no better option in case of U-Boot unfortunatelly. This change will
most likely increase the duration of 'usb start' on some platforms,
but this is needed to fix a bug.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Cc: Hans de Goede <hdegoede@redhat.com>
9 years agonet: asix: fix operation without eeprom
Marcel Ziswiler [Wed, 5 Aug 2015 15:16:59 +0000 (17:16 +0200)]
net: asix: fix operation without eeprom

This patch fixes operation of our on-board AX88772B chip without EEPROM
but with a ethaddr coming from the regular U-Boot environment. This is
a forward port of some remaining parts initially implemented by
Antmicro.

Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Marek Vasut <marex@denx.de>
9 years agosunxi: Drop our own copy of the USB_KEYBOARD options
Hans de Goede [Tue, 4 Aug 2015 15:04:13 +0000 (17:04 +0200)]
sunxi: Drop our own copy of the USB_KEYBOARD options

USB_KEYBOARD is now defined in drivers/usb/Kconfig, drop our own duplicate
definition.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
9 years agousb: USB storage-specific part ifdef in uclass
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:12 +0000 (17:04 +0200)]
usb: USB storage-specific part ifdef in uclass

usb_stor_reset is only defined when USB storage support is enabled, thus the
function is not declared when such support is missing.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agosunxi: MUSB gadget config provisions, with fastboot and USB mass storage support
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:11 +0000 (17:04 +0200)]
sunxi: MUSB gadget config provisions, with fastboot and USB mass storage support

This adds some config options to the sunxi-common config to enable the USB
download gadget and the fastboot and USB mass storage functions.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agomusb: sunxi: Force EP0 on re-enable
Maxime Ripard [Tue, 4 Aug 2015 15:04:10 +0000 (17:04 +0200)]
musb: sunxi: Force EP0 on re-enable

Currently, the second use of a gadget will fail, while the first one works.

Forcing the EP0 at every enable fix this issue.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agosunxi: config update to stick with Kconfig changes
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:09 +0000 (17:04 +0200)]
sunxi: config update to stick with Kconfig changes

Now that some things were updated in Kconfig, they should be reflected on the
sunxi-common config too.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agosunxi: usb: USB_MUSB_SUNXI move to musb-new Kconfig
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:08 +0000 (17:04 +0200)]
sunxi: usb: USB_MUSB_SUNXI move to musb-new Kconfig

Now that the musb-new driver has a Kconfig, we can move Kconfig options to
enable controllers to it, so that it's easier in e.g. menuconfig.

In addition, this allows declaring support for USB_MUSB_HOST/GADGET in
defconfigs instead of the USB_MUSB_SUNXI controller, that will get selected
automatically when needed.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: musb-new: Kconfig support for USB_MUSB_HOST and USB_MUSB_GADGET
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:07 +0000 (17:04 +0200)]
usb: musb-new: Kconfig support for USB_MUSB_HOST and USB_MUSB_GADGET

Having MUSB_HOST and MUSB_GADGET in Kconfig allows more flexibility with regard
to what Kconfig options to enable, such as USB_STORAGE or USB_KEYBOARD.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: musb-new: CONFIG_MUSB prefix replacement with CONFIG_USB_MUSB
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:06 +0000 (17:04 +0200)]
usb: musb-new: CONFIG_MUSB prefix replacement with CONFIG_USB_MUSB

USB-related options are usually prefixed with CONFIG_USB and platform-specific
adaptation for the MUSB controller already have a CONFIG_USB_MUSB prefix, so
this switches all MUSB-related options to a CONFIG_USB_MUSB prefix, for
consistency.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: Generic USB Kconfig option, that fits both host and gadget and comments
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:05 +0000 (17:04 +0200)]
usb: Generic USB Kconfig option, that fits both host and gadget and comments

There is no particular reason why the USB Kconfig option should be specific to
host mode. In prevision of adding MUSB host and gadget to Kconfig, this moves
the title and help message of the USB Kconfig option to a more generic format.

Adding comments to the usb Kconfig allows for a better separation and more
readability in generated configs and in menuconfig.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: USB_ARCH_HAS_HCD Kconfig option removal
Paul Kocialkowski [Tue, 4 Aug 2015 15:04:04 +0000 (17:04 +0200)]
usb: USB_ARCH_HAS_HCD Kconfig option removal

The USB_ARCH_HAS_HCD currently serves no purpose and adds some confusion to the
required Kconfig options that are required to have USB support.

Dropping it makes things easier and doesn't break anything, since it was unused
anyways.

Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
9 years agousb: dwc2: Call board_usb_init() from usb_lowlevel_init()
Peter Griffin [Wed, 29 Jul 2015 21:39:29 +0000 (22:39 +0100)]
usb: dwc2: Call board_usb_init() from usb_lowlevel_init()

This patch makes the dwc2 controller like ehci / ohci / xhci controllers
by calling the board_usb_init() function from usb_lowlevel_init.

This can then be implemented by specific platforms to initialise
their USB hardware (phys / clocks etc).

Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
9 years agoefi: Add a README to explain how things work
Simon Glass [Tue, 4 Aug 2015 18:34:05 +0000 (12:34 -0600)]
efi: Add a README to explain how things work

Add some documentation on the EFI implementation in U-Boot.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Gracefully disable the vesa driver when running from EFI
Simon Glass [Tue, 4 Aug 2015 18:34:04 +0000 (12:34 -0600)]
x86: Gracefully disable the vesa driver when running from EFI

We cannot use this driver when running from EFI as we have no direct hardware
access. In fact coreboot uses a different driver which uses tables provided
by coreboot. So far it does not seem possible to use a normal video driver
when booting from EFI.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Anatolij Gustschin <agust@denx.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: qemu: Support operation as an EFI payload
Simon Glass [Tue, 4 Aug 2015 18:34:03 +0000 (12:34 -0600)]
x86: qemu: Support operation as an EFI payload

Disable a few things which interfere with the EFI init. This allows QEMU to
to boot into EFI, load a U-Boot payload then boot to the U-Boot prompt.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: baytrail: Support operation as an EFI payload
Simon Glass [Tue, 4 Aug 2015 18:34:02 +0000 (12:34 -0600)]
x86: baytrail: Support operation as an EFI payload

Disable a few things which interfere with the EFI init. This allows the
Minnowboard MAX to boot into EFI, load a U-Boot payload then boot to the
U-Boot prompt.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add helper code for running from EFI
Simon Glass [Tue, 4 Aug 2015 18:34:01 +0000 (12:34 -0600)]
x86: Add helper code for running from EFI

When U-Boot is running from EFI some of the x86 init is replaced with
EFI-specific init. For example, since DRAM has already been set up, we only
need to find it, not init it. Add these functions so that boards can easily
allow booting from EFI if required.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Handle running as EFI payload
Simon Glass [Tue, 4 Aug 2015 18:34:00 +0000 (12:34 -0600)]
x86: Handle running as EFI payload

When U-Boot runs as an EFI payload it needs to avoid setting up the CPU
again. Also U-Boot currently does not handle interrupts for many devices, so
run with interrupts disabled.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Add a command to display the memory map
Simon Glass [Tue, 4 Aug 2015 18:33:59 +0000 (12:33 -0600)]
efi: Add a command to display the memory map

The EFI memory map is passed from the stub to U-Boot in a table. Add a
command to display it in a vaguely readable fashion.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested on QEMU
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Add functions for decoding the EFI tables
Simon Glass [Tue, 4 Aug 2015 18:33:58 +0000 (12:33 -0600)]
efi: Add functions for decoding the EFI tables

The EFI stub can pass a table to U-Boot with information about the memory map
Potentially other things will follow. Add a way to access this table.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add support for passing tables into U-Boot
Simon Glass [Tue, 4 Aug 2015 18:33:57 +0000 (12:33 -0600)]
x86: Add support for passing tables into U-Boot

The EFI stub provides information to U-Boot in a table. This includes the
memory map which is needed to decide where to relocate U-Boot. Collect this
information in the early init code and store it in global_data.

Fix up the BIST code at the same time since we don't have it when booting
from EFI and can assume it is 0.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Add 64-bit payload support
Simon Glass [Tue, 4 Aug 2015 18:33:56 +0000 (12:33 -0600)]
efi: Add 64-bit payload support

Most EFI implementations use 64-bit. Add a way to build U-Boot as a 64-bit
EFI payload. The payload unpacks a (32-bit) U-Boot and starts it. This can
be enabled for x86 boards at present.

Signed-off-by: Simon Glass <sjg@chromium.org>
Improvements to how the payload is built:
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add a way to call 32-bit code from 64-bit mode
Simon Glass [Tue, 4 Aug 2015 18:33:55 +0000 (12:33 -0600)]
x86: Add a way to call 32-bit code from 64-bit mode

The procedure to drop from 64-bit mode to 32-bit is a bit messy. Add a
function to take care of it. It requires identity-mapped pages and that
the calling code is running below 4GB.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add an enum for some commonly-used GDT bits
Simon Glass [Tue, 4 Aug 2015 18:33:54 +0000 (12:33 -0600)]
x86: Add an enum for some commonly-used GDT bits

Rather than add these as open-coded values, create an enum with the commonly
used flags.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Support building the EFI stub
Simon Glass [Tue, 4 Aug 2015 18:33:53 +0000 (12:33 -0600)]
x86: Support building the EFI stub

Add support for building a 32/64-bit EFI stub for x86. This involves
building the startup and relocation code for either i386 or x86_64.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Add support for loading U-Boot through an EFI stub
Simon Glass [Tue, 4 Aug 2015 18:33:52 +0000 (12:33 -0600)]
efi: Add support for loading U-Boot through an EFI stub

It is useful to be able to load U-Boot onto a board even if is it already
running EFI. This can allow access to the U-Boot command interface, flexible
booting options and easier development.

The easiest way to do this is to build U-Boot as a binary blob and have an
EFI stub copy it into RAM. Add support for this feature, targeting 32-bit
initially.

Also add a way to detect when U-Boot has been loaded via a stub. This goes
in common.h since it needs to be widely available so that we avoid redoing
initialisation that should be skipped.

Signed-off-by: Simon Glass <sjg@chromium.org>
Improvements to how the payload is built:
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add relocation and link script for a 64-bit EFI application
Simon Glass [Tue, 4 Aug 2015 18:33:51 +0000 (12:33 -0600)]
x86: Add relocation and link script for a 64-bit EFI application

Add a linker script and relocation code for building 64-bit EFI
applications. This can be used for the EFI stub.

Signed-off-by: Simon Glass <sjg@chromium.org>
Improvements to how the payload is built:
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add definitions for the x86-efi board and plumb it in
Ben Stoltz [Tue, 4 Aug 2015 18:33:50 +0000 (12:33 -0600)]
x86: Add definitions for the x86-efi board and plumb it in

Add configuration and Kconfig changes for this board.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Allow relocation code to build without text base
Simon Glass [Tue, 4 Aug 2015 18:33:49 +0000 (12:33 -0600)]
x86: Allow relocation code to build without text base

This code currently requires CONFIG_SYS_TEXT_BASE but this should be
unnecessary. As a first step, remove the build-time limitation and report an
error instead.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: dts: Add a device tree file for EFI
Simon Glass [Tue, 4 Aug 2015 18:33:48 +0000 (12:33 -0600)]
x86: dts: Add a device tree file for EFI

This contains just enough to bring up the serial UART.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add EFI board code
Ben Stoltz [Tue, 4 Aug 2015 18:33:47 +0000 (12:33 -0600)]
x86: Add EFI board code

Add support for the efi-x86 board, which supports running U-Boot as an
EFI 32-bit application.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add support for U-Boot as an EFI application
Ben Stoltz [Tue, 4 Aug 2015 18:33:46 +0000 (12:33 -0600)]
x86: Add support for U-Boot as an EFI application

Add the required x86 glue code. This includes the initial start-up,
relocation and jumping to efi_main(). We also need to avoid fiddling with
interrupts.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add asm/elf.h for x86-specific ELF definitions
Simon Glass [Tue, 4 Aug 2015 18:33:45 +0000 (12:33 -0600)]
x86: Add asm/elf.h for x86-specific ELF definitions

Bring in this file from Linux 4.1. It supports relocation features specific
to x86.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Support skipping relocation for EFI
Simon Glass [Tue, 4 Aug 2015 18:33:44 +0000 (12:33 -0600)]
x86: Support skipping relocation for EFI

When running as an EFI application we must skip relocation. Add support for
this in the x86 relocation code.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Support building a u-boot-app.efi executable
Simon Glass [Tue, 4 Aug 2015 18:33:43 +0000 (12:33 -0600)]
efi: Support building a u-boot-app.efi executable

Add support for building U-Boot as an EFI application with a .efi suffix.
This can be loaded by EFI provided that EFI has the same bit width (32-
or 64-bit) as U-Boot. This unfortunate limitation is imposed by EFI.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Set up toolchain flags for running as EFI application
Ben Stoltz [Tue, 4 Aug 2015 18:33:42 +0000 (12:33 -0600)]
x86: Set up toolchain flags for running as EFI application

Adjust the toolchain flags to build U-Boot as a relocatable shared library,
as required by EFI.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Drop CONFIG_SYS_TEXT_BASE for EFI
Ben Stoltz [Tue, 4 Aug 2015 18:33:41 +0000 (12:33 -0600)]
efi: Drop CONFIG_SYS_TEXT_BASE for EFI

When U-Boot runs as an EFI application is does not have a definition of
CONFIG_SYS_TEXT_BASE. U-Boot is a relocatable application and the relocation
is done by EFI. U-Boot can be loaded at any address.

Ensure that this CONFIG option is not set in this case.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Add a serial driver
Simon Glass [Tue, 4 Aug 2015 18:33:40 +0000 (12:33 -0600)]
efi: Add a serial driver

Add a serial driver which makes use of EFI's console in/out service.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested on Intel Crown Bay and QEMU
Tested-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoAdd a way to skip relocation
Simon Glass [Tue, 4 Aug 2015 18:33:39 +0000 (12:33 -0600)]
Add a way to skip relocation

When running U-Boot as an EFI application we cannot relocate since we do not
have relocation information. U-Boot has already been relocated to a suitable
address.

Add a global_data flag to control skipping relocation.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Update README to explain booting Ubuntu on Minnowmax
Simon Glass [Mon, 27 Jul 2015 21:47:31 +0000 (15:47 -0600)]
x86: Update README to explain booting Ubuntu on Minnowmax

The steps required to boot a Linux distribution from U-Boot on x86 are not
very complicated, but it is a good idea to have these written down in an
accessible place.

Document how to examine the boot media from U-Boot, how to load a kernel,
load a ramdisk, set the kernel boot arguments and start the kernel. With
these instructions Ubuntu boots mostly normally on Minnowmax.

Note that the TSC timer does not operate correctly and gives warnings in
the boot log. I expect that ACPI support will solve this.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: minnowmax: Drop the old PCI settings
Simon Glass [Mon, 27 Jul 2015 21:47:29 +0000 (15:47 -0600)]
x86: minnowmax: Drop the old PCI settings

These are now in the device tree so we don't need to use the CONFIG options.
Drop them.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: dts: Fix typo in intel,irq-router.txt
Simon Glass [Mon, 27 Jul 2015 21:47:26 +0000 (15:47 -0600)]
x86: dts: Fix typo in intel,irq-router.txt

Fix a small typo in this binding file.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Drop CONFIG_DCACHE_RAM_MRC_VAR_SIZE from header file
Simon Glass [Mon, 27 Jul 2015 21:47:25 +0000 (15:47 -0600)]
x86: Drop CONFIG_DCACHE_RAM_MRC_VAR_SIZE from header file

This is provided in Kconfig so we don't need it here.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Move CONFIG_X86_SERIAL to Kconfig
Simon Glass [Mon, 27 Jul 2015 21:47:23 +0000 (15:47 -0600)]
x86: Move CONFIG_X86_SERIAL to Kconfig

Move this config option to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Move Chrome OS options to defconfig
Simon Glass [Mon, 27 Jul 2015 21:47:22 +0000 (15:47 -0600)]
x86: Move Chrome OS options to defconfig

Drop these from the header file and use Kconfig instead.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agodm: Add a return value comment to device_get_child()
Simon Glass [Mon, 27 Jul 2015 21:47:19 +0000 (15:47 -0600)]
dm: Add a return value comment to device_get_child()

At present this function does not specify its return value. Fix it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agopci: Indicate prefetchable memory allocate when debugging
Simon Glass [Mon, 27 Jul 2015 21:47:18 +0000 (15:47 -0600)]
pci: Indicate prefetchable memory allocate when debugging

At present the PCI output displays 'Mem' when it allocates memory for a PCI
device, whether it is prefetchable or not. There is a distinction since the
memory comes from separate pools. Use 'Prf' instead of 'Mem' when allocating
prefetchable memory.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agopci: Add a constant for an invalid interrupt
Simon Glass [Mon, 27 Jul 2015 21:47:17 +0000 (15:47 -0600)]
pci: Add a constant for an invalid interrupt

Rather than using 0xff in the code, add a constant.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Display the correct initcall pre-relocation values
Simon Glass [Fri, 31 Jul 2015 15:31:38 +0000 (09:31 -0600)]
efi: Display the correct initcall pre-relocation values

With EFI the start address of U-Boot is specified differently. We could
consider just setting GD_FLG_RELOC and then setting up reloc_off. But that
flag has other implementations and we are not able to use U-Boot relocation
which this flag implies.

Instead, just add a special case for EFI.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Avoid using non-existent text base
Ben Stoltz [Fri, 31 Jul 2015 15:31:37 +0000 (09:31 -0600)]
efi: Avoid using non-existent text base

When U-Boot runs as an EFI application is does not have a definition of
CONFIG_SYS_TEXT_BASE. U-Boot is a relocatable application and the relocation
is done by EFI. U-Boot can be loaded at any address.

This is similar to how sandbox works. Adjust the early board init to deal
with this.

Signed-off-by: Ben Stoltz <stoltz@google.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoefi: Add start-up library code
Simon Glass [Fri, 31 Jul 2015 15:31:36 +0000 (09:31 -0600)]
efi: Add start-up library code

When running as an EFI application, U-Boot must request memory from EFI,
and provide access to the boot services U-Boot needs.

Add library code to perform these tasks. This includes efi_main() which is
the entry point from EFI. U-Boot is built as a shared library.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Allow use of global_data with EFI
Simon Glass [Fri, 31 Jul 2015 15:31:35 +0000 (09:31 -0600)]
x86: Allow use of global_data with EFI

On x86 the global_data pointer is provided through a somewhat-bizarre and
x86-specific mechanism: the F segment register is set to a pointer to the
start of global_data, so that accesses can use this build-in register.

When running as an EFI application we don't want to mess with the Global
Descriptor Table (GDT) and there is little advantage (in terms of code size)
to doing so.

Allow global_data to be a simple variable in this case.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agopci: Fix up code for CONFIG_PCI_ENUM_ONLY
Simon Glass [Fri, 31 Jul 2015 15:31:34 +0000 (09:31 -0600)]
pci: Fix up code for CONFIG_PCI_ENUM_ONLY

This option is not used by any board but appears to still be useful, at least
for testing. With recent commits it does not build, so fix it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agopci: Use debug() instead of DEBUGF() in pci_auto.c
Simon Glass [Fri, 31 Jul 2015 15:31:33 +0000 (09:31 -0600)]
pci: Use debug() instead of DEBUGF() in pci_auto.c

This file does not need its own way of doing debug(). Clean it up to use the
new way.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Pavel Machek <pavel@denx.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Tidy up a few minor issues with interrupts
Simon Glass [Fri, 31 Jul 2015 15:31:32 +0000 (09:31 -0600)]
x86: Tidy up a few minor issues with interrupts

Fix a typo, remove an unused field and make sure to use existing #define
constants instead of open-coded values.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add some missing global_data declarations in files that use gd
Simon Glass [Fri, 31 Jul 2015 15:31:31 +0000 (09:31 -0600)]
x86: Add some missing global_data declarations in files that use gd

Some files use global_data but don't declare it. Fix this.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Tidy up the 64-bit calling code
Simon Glass [Fri, 31 Jul 2015 15:31:30 +0000 (09:31 -0600)]
x86: Tidy up the 64-bit calling code

The GDT works but technically the length is incorrect. Fix this and add a
comment.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Drop unused copy_fdt_to_ram()
Simon Glass [Fri, 31 Jul 2015 15:31:29 +0000 (09:31 -0600)]
x86: Drop unused copy_fdt_to_ram()

This is now handled by generic U-Boot code so we do not need an x86 version.
It is no-longer called, so remove it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Tidy up global_data flags
Simon Glass [Fri, 31 Jul 2015 15:31:28 +0000 (09:31 -0600)]
x86: Tidy up global_data flags

These flags now overlap some global ones. Adjust the x86-specific flags to
avoid this. Since this requires a change to the start.S code, add a way for
tools to find the 32-bit cold reset entry point. Previously this was at a
fixed offset.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoReserve the top 16 flag bits for architecture-specific use
Simon Glass [Fri, 31 Jul 2015 15:31:27 +0000 (09:31 -0600)]
Reserve the top 16 flag bits for architecture-specific use

Add a convention that the generic global_data only occupy the bottom 16 bits
of the flags word, so that there is less chance of a conflict. At present the
x86 flags conflict.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Use CR0 constants in CPU init
Simon Glass [Fri, 31 Jul 2015 15:31:26 +0000 (09:31 -0600)]
x86: Use CR0 constants in CPU init

We should use these constants where possible.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agox86: Add various minor tidy-ups to the 32-bit startup code
Simon Glass [Fri, 31 Jul 2015 15:31:25 +0000 (09:31 -0600)]
x86: Add various minor tidy-ups to the 32-bit startup code

Fix a typo, improve some comments and add a little more detail in some
cases.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoBring in __aligned_u64 and friends to linux/types.h
Simon Glass [Fri, 31 Jul 2015 15:31:24 +0000 (09:31 -0600)]
Bring in __aligned_u64 and friends to linux/types.h

These will be used for efi.h both for U-Boot running as an EFI application
and as a payload. They come from Linux 4.1.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoAllow objcopy to work without filling gaps with 0xff
Simon Glass [Fri, 31 Jul 2015 15:31:23 +0000 (09:31 -0600)]
Allow objcopy to work without filling gaps with 0xff

This is currently done for all targets, since 0xff is the default erased
value for most flash devices. In some cases this is not what we want (e.g.
for EFI images) so provide a command to do a vanilla objcopy.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoelf: Add a few definitions for 64-bit relocation
Simon Glass [Fri, 31 Jul 2015 15:31:21 +0000 (09:31 -0600)]
elf: Add a few definitions for 64-bit relocation

Provide the types necessary to relocate 64-bit images.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agodm: serial: Correct logic in serial_find_console_or_panic()
Simon Glass [Fri, 31 Jul 2015 15:31:20 +0000 (09:31 -0600)]
dm: serial: Correct logic in serial_find_console_or_panic()

In a final attempt to find a console UART this function uses the first
first available serial device. However the check for a valid device is
inverted.

This code is only executed when there is in fact no serial UART, but at
present it can fail to reach the panic_str() call in this case, and start
trying to use a non-existent UART.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agodm: core: Check for empty list in uclass_find_device()
Simon Glass [Fri, 31 Jul 2015 15:31:19 +0000 (09:31 -0600)]
dm: core: Check for empty list in uclass_find_device()

This function needs to check the list has entries before traversing it.
Fix this bug.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
9 years agoSupport removing default assembler flags
Simon Glass [Fri, 31 Jul 2015 15:31:18 +0000 (09:31 -0600)]
Support removing default assembler flags

The CFLAGS_REMOVE_<file> feature allows default C compiler flags to be
removed for particular files. Add the same feature for assembler, using
AFLAGS_REMOVE_<file>.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
9 years agocommon: command add '\n' for debug msg
Peng Fan [Tue, 28 Jul 2015 14:45:37 +0000 (22:45 +0800)]
common: command add '\n' for debug msg

Add '\n' for debug msg.

Signed-off-by: Peng Fan <Peng.Fan@freescale.com>
Cc: Tom Rini <trini@konsulko.com>
Cc: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agolib: fdt: fix indent of #ifdef..#endif conditional
Masahiro Yamada [Sat, 1 Aug 2015 07:03:25 +0000 (16:03 +0900)]
lib: fdt: fix indent of #ifdef..#endif conditional

Match the depth of indentation between #ifdef and #endif
for better readability.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agocommon: Print nothing in the __weak checkboard()
Bin Meng [Fri, 31 Jul 2015 07:24:40 +0000 (00:24 -0700)]
common: Print nothing in the __weak checkboard()

Do not print confusing "Board: Unknown" during boot.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agogitignore: Add defconfig and fdtgrep
Bin Meng [Fri, 31 Jul 2015 07:24:39 +0000 (00:24 -0700)]
gitignore: Add defconfig and fdtgrep

Ignore defconfig and tools/fdtgrep.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: bayleybay: Configure PCI IRQ
Bin Meng [Thu, 30 Jul 2015 10:49:18 +0000 (03:49 -0700)]
x86: bayleybay: Configure PCI IRQ

Add PCI IRQ routing information in the board device tree and enable
writing PIRQ routing table and MP table.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: Add Intel Bayley Bay board support
Bin Meng [Thu, 30 Jul 2015 10:49:17 +0000 (03:49 -0700)]
x86: Add Intel Bayley Bay board support

Intel Bayley Bay board is a BayTrail based board. Add this board
with existing baytrail fsp support.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: Add microcode for BayTrail-I B0 stepping
Bin Meng [Thu, 30 Jul 2015 10:49:16 +0000 (03:49 -0700)]
x86: Add microcode for BayTrail-I B0 stepping

This commit adds the microcode blob for BayTrail-I B0 stepping,
CPUID signature 30671h.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: baytrail: Update UPD setting for FSP Gold4 release
Bin Meng [Thu, 30 Jul 2015 10:49:15 +0000 (03:49 -0700)]
x86: baytrail: Update UPD setting for FSP Gold4 release

BayTrail FSP Gold4 release adds one UPD parameter to control IGD
enable/disable.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: minnowmax: Remove smsc47x superio codes
Bin Meng [Thu, 30 Jul 2015 10:49:14 +0000 (03:49 -0700)]
x86: minnowmax: Remove smsc47x superio codes

On Intel BayTrail SoC, there is a legacy UART (I/O 0x3f8) integrated
into the SoC which is enabled by the FSP. Remove the smsc47x superio
initialization codes.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: Sanity test on vesa parameters before setting up kernel screen_info
Bin Meng [Thu, 30 Jul 2015 10:49:13 +0000 (03:49 -0700)]
x86: Sanity test on vesa parameters before setting up kernel screen_info

We should not set up kernel screen_info when the vesa parameters are
insane, otherwise kernel will panic.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agox86: qemu: Add MP initialization
Bin Meng [Mon, 27 Jul 2015 11:16:08 +0000 (19:16 +0800)]
x86: qemu: Add MP initialization

Add a cpu1 node to the device tree and enable the MP initialization
on QEMU targets (i440fx and q35).

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
9 years agox86: Add a 'pause' instruction in __udelay() for QEMU target
Miao Yan [Mon, 27 Jul 2015 11:16:07 +0000 (19:16 +0800)]
x86: Add a 'pause' instruction in __udelay() for QEMU target

When running SMP configuration on QEMU (tcg mode, no kvm), there is
a busy loop in start_aps(), calling udelay(), that waits for APs to
show up online. However, there is a chance that VCPU1 will be timeout
waiting, IOW the secondary VCPUs haven't started their execution yet.

This patch adds a 'pause' instruction in __udelay() only for QEMU
target, to give other VCPUs a chance to run. When QEMU sees the
'pause' instruction, it will yeild the execution to other CPUs.

Signed-off-by: Miao Yan <yanmiaobest@gmail.com>
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
9 years agodoc: dm: Update pci-info.txt for pci support
Bin Meng [Mon, 27 Jul 2015 07:33:43 +0000 (00:33 -0700)]
doc: dm: Update pci-info.txt for pci support

Correct two typos and mention how pci bus will be probed.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: pci: Allow scan bridge child devices before relocation
Bin Meng [Mon, 27 Jul 2015 07:33:39 +0000 (00:33 -0700)]
dm: pci: Allow scan bridge child devices before relocation

On some platforms pci devices behind bridge need to be probed (eg:
a pci uart on recent x86 chipset) before relocation. Remove such
limitation so that dm pci can be used before relocation.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-fsl-qoriq
Tom Rini [Tue, 4 Aug 2015 16:21:45 +0000 (12:21 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-fsl-qoriq

9 years agoconfigs: ti_armv7_keystone2: replace addr_uboot with loadaddr
Vitaly Andrianov [Mon, 3 Aug 2015 19:54:32 +0000 (15:54 -0400)]
configs: ti_armv7_keystone2: replace addr_uboot with loadaddr

This patch replaces not existing addr_uboot environment variable by
loadaddr at get_uboot_net and burn_uboot_xxx commands. Otherwise these
commands are broken.

Signed-off-by: Vitaly Andrianov <vitalya@ti.com>
Acked-by: Nishanth Menon <nm@ti.com>
9 years agoarmv8/ls2085ardb: Add USB support for ls2085ardb
Nikhil Badola [Fri, 26 Jun 2015 11:32:18 +0000 (17:02 +0530)]
armv8/ls2085ardb: Add USB support for ls2085ardb

Add USB XHCI support for ls2085rdb platform

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoarmv8/ls2085aqds: Add USB support for ls2085aqds
Nikhil Badola [Fri, 26 Jun 2015 11:32:00 +0000 (17:02 +0530)]
armv8/ls2085aqds: Add USB support for ls2085aqds

Add USB XHCI support for ls2085qds platform

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoarmv8/lsch3/config: Define USB XHCI controller base address for LS2085A
Nikhil Badola [Fri, 26 Jun 2015 11:31:50 +0000 (17:01 +0530)]
armv8/lsch3/config: Define USB XHCI controller base address for LS2085A

Define base address of both usb xhci controllers in lsch3 config
in the format (IMMR + offset) for LS2085A

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoarmv8/lsch3/config: Define CONFIG_SYS_CACHELINE_SIZE for LS2085A
Nikhil Badola [Fri, 26 Jun 2015 11:29:21 +0000 (16:59 +0530)]
armv8/lsch3/config: Define CONFIG_SYS_CACHELINE_SIZE for LS2085A

Define CONFIG_SYS_CACHELINE_SIZE for LS2085A which is required by
USB XHCI stack for alignment

Signed-off-by: Nikhil Badola <nikhil.badola@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agodrivers/ddr/fsl: Adjust bstopre value
York Sun [Thu, 23 Jul 2015 21:04:48 +0000 (14:04 -0700)]
drivers/ddr/fsl: Adjust bstopre value

By default the bstopre value has been set to 0x100, used to be 1/4
value of refint. Modern DDR has increased the refresh time. Adjust
to 1/4 of refresh interval dynamically. Individual board can still
override this value in board ddr file, or to use auto-precharge.

Signed-off-by: York Sun <yorksun@freescale.com>
9 years agodrivers/crypto/fsl: clean-up - use fdt_setprop_u32 helper
horia.geanta@freescale.com [Wed, 8 Jul 2015 14:24:58 +0000 (17:24 +0300)]
drivers/crypto/fsl: clean-up - use fdt_setprop_u32 helper

Signed-off-by: Horia Geantă <horia.geanta@freescale.com>
Acked-by: Ruchika Gupta<ruchika.gupta@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agodrivers/crypto/fsl: fix snooping for write transactions
horia.geanta@freescale.com [Wed, 8 Jul 2015 14:24:57 +0000 (17:24 +0300)]
drivers/crypto/fsl: fix snooping for write transactions

HW coherency won't work properly for CAAM write transactions
if AWCACHE is left to default (POR) value - 4'b0001.
It has to be programmed to 4'b0010.

For platforms that have HW coherency support:
-PPC-based: the update has no effect; CAAM coherency already works
due to the IOMMU (PAMU) driver setting the correct memory coherency
attributes
-ARM-based: the update fixes cache coherency issues,
since IOMMU (SMMU) driver is not programmed to behave similar to PAMU

Fixes: b9eebfade974c ("fsl_sec: Add hardware accelerated SHA256 and SHA1")
Signed-off-by: Horia Geantă <horia.geanta@freescale.com>
Reviewed-by: Aneesh Bansal <aneesh.bansal@freescale.com>
Reviewed-by: Mingkai Hu <Mingkai.Hu@freescale.com>
Acked-by: Ruchika Gupta<ruchika.gupta@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>