chenhui zhao [Fri, 23 Jan 2015 07:53:53 +0000 (15:53 +0800)]
arm: ls102xa: workaround for cache coherency problem
The RCPM FSM may not be reset after power-on, for example,
in the cases of cold boot and wakeup from deep sleep.
It causes cache coherency problem and may block deep sleep.
Therefore, reset them if they are not be reset.
Signed-off-by: Chenhui Zhao <chenhui.zhao@freescale.com> Reviewed-by: York Sun <yorksun@freescale.com>
Minghuan Lian [Wed, 21 Jan 2015 09:29:17 +0000 (17:29 +0800)]
arm/ls102xa: create TLB to map PCIe region
LS1021A's PCIe1 region begins 0x40_00000000; PCIe2 begins
0x48_00000000. In order to access PCIe device, we must create
TLB to map the 40bit physical address to 32bit virtual address.
This patch will enable MMU after DDR is available and creates MMU
table in DRAM to map all 4G space; then, re-use the reserved space
to map PCIe region. The following the mapping layout.
VA mapping:
------- <---- 0GB
| |
| |
|-------| <---- 0x24000000
|///////| ===> 192MB VA map for PCIe1 with offset 0x40_0000_0000
|-------| <---- 0x300000000
| |
|-------| <---- 0x34000000
|///////| ===> 192MB VA map for PCIe2 with offset 0x48_0000_0000
|-------| <---- 0x40000000
| |
|-------| <---- 0x80000000 DDR0 space start
|\\\\\\\|
|\\\\\\\| ===> 2GB VA map for 2GB DDR0 Memory space
|\\\\\\\|
------- <---- 4GB DDR0 space end
Signed-off-by: Minghuan Lian <Minghuan.Lian@freescale.com> Reviewed-by: York Sun <yorksun@freescale.com>
Upgrade Manage Complex (MC) flib API to 0.5.2. Rename directory
fsl_mc to fsl-mc. Change the fsl-mc node in Linux device tree
from "fsl,dprcr" to "fsl-mc". Print MC version info when
appropriate.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com> Signed-off-by: Lijun Pan <Lijun.Pan@freescale.com> Reviewed-by: York Sun <yorksun@freescale.com>
Kuldip Giroh [Tue, 6 Jan 2015 21:18:53 +0000 (13:18 -0800)]
ARMv8/LS2085A: HugeTLB support is required by default in LS NADK
LS NADK memory manager by default works on HugeTLB. Hence bootargs
must include parameters default_hugepagesz (default hugepagesize,
hugepagesz (hugepage size) and hugepages (number of hugepages to be
reserved in kernel for the given size).
Signed-off-by: Kuldip Giroh <kuldip.giroh@freescale.com> Reviewed-by: York Sun <yorksun@freescale.com>
York Sun [Tue, 6 Jan 2015 21:18:50 +0000 (13:18 -0800)]
driver/ddr/fsl: Add support for multiple DDR clocks
Controller number is passed for function calls to support individual
DDR clock, depending on SoC implementation. It is backward compatible
with exising platforms. Multiple clocks have been verifyed on LS2085A
emulator.
York Sun [Tue, 6 Jan 2015 21:18:49 +0000 (13:18 -0800)]
armv8/fsl-lsch3: Add support for second DDR clock
FSL-LSCH3 platforms can have multiple DDR clocks. LS2085A has one clock for
general DDR controlers, and another clock for DP-DDR. DDR driver needs to
change to support multiple clocks.
York Sun [Tue, 6 Jan 2015 21:18:42 +0000 (13:18 -0800)]
armv8/fsl-lsch3: Convert flushing L3 to assembly to avoid using stack
Flushing L3 cache in CCN-504 requries d-cache to be disabled. Using
assembly function to guarantee stack is not used before flushing is
completed. Timeout is needed for simualtor on which CCN-504 is not
implemented. Return value can be checked for timeout situation.
Change bootm.c to disable dcache instead of simply flushing, required
by flushing L3.
Bhupesh Sharma [Tue, 6 Jan 2015 21:11:21 +0000 (13:11 -0800)]
fsl-ch3/lowlevel: TZPC and TZASC programming to configure non-secure accesses
This patch ensures that the TZPC (BP147) and TZASC-400 programming
happens for LS2085A SoC only when the desired config flags are
enabled and ensures that the TZPC programming is done to allow Non-secure
(NS) + secure (S) transactions only for DCGF registers.
The TZASC component is not present on LS2085A-Rev1, so the TZASC-400
config flag is turned OFF for now.
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@freescale.com> Reviewed-by: York Sun <yorksun@freescale.com>
Matt Reimer [Thu, 19 Feb 2015 18:22:53 +0000 (11:22 -0700)]
mmc: sdhci: fix bus width switching on Samsung SoCs
Fix bus width switching from 8-bit mode down to 4-bit or 1-bit modes on
Samsung SoCs using SDHCI_QUIRK_USE_WIDE8. These SoCs report controller
version 2.0 yet they support 8-bit bus widths. If 8-bit mode was
previously enabled and then an operation like "mmc dev" caused a switch
back down to 4-bit or 1-bit mode, WIDE8 was left set, causing failures.
This problem was manifested by "mmc dev" timing out.
Signed-off-by: Matt Reimer <mreimer@sdgsystems.com>
mmc: exynos dwmmc: check boot mode before init dwmmc
Before this commit, the mmc devices were always registered
in the same order. So dwmmc channel 0 was registered as mmc 0,
channel 1 as mmc 1, etc.
In case of possibility to boot from more then one device,
the CONFIG_SYS_MMC_ENV_DEV should always point to right mmc device.
This can be achieved by init boot device as first, so it will be
always registered as mmc 0. Thanks to this, the 'saveenv' command
will work fine for all mmc boot devices.
Exynos based boards usually uses mmc host channels configuration:
- 0, or 0+1 for 8 bit - as a default boot device (usually eMMC)
- 2 for 4bit - as an optional boot device (usually SD card slot)
And usually the boot order is defined by OM pin configuration,
which can be changed in a few ways, eg.
- Odroid U3 - eMMC card insertion -> first boot from eMMC
- Odroid X2/XU3 - boot priority jumper
By this commit, Exynos dwmmc driver will check the OM pin configuration,
and then try to init the boot device and register it as mmc 0.
Hans de Goede [Thu, 19 Feb 2015 20:03:21 +0000 (21:03 +0100)]
sunxi: mmc: Always declare High Capacity capability
High Capacity (e)MMC cards work fine on sun4i / sun5i, and not having this
capability set causes u-boot to not recognize the eMMC on an Utoo P66 A13
tablet, so always set it thereby fixing this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Ian Campbell <ijc@hellion.org.uk>
Jaehoon Chung [Wed, 4 Feb 2015 06:48:40 +0000 (15:48 +0900)]
mmc: exynos_dw_mmc: use the exynos specific data structure
Clksel value is exynos specific value.
It removed "clksel_val" into dwmci_host and created the
"dwmci_exynos_priv_data" structure for exynos specific data.
Jaehoon Chung [Wed, 4 Feb 2015 06:48:39 +0000 (15:48 +0900)]
mmc: exynos_dw_mmc: set to clksel_val into board-init function
"clksel_val" is assigned to property of mmc or defined value.
But it doesn't write at initial sequence.
There is a reason that get the wrong source-clock value.
This patch fixed it.
The SD/MMC version scheme was buggy when dealing with standard
major.minor.change cases. Fix it by using something similar to
the linux's kernel versioning method.
Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com> Tested-by: Jaehoon Chung <jh80.chung@samsung.com> Reported-by: Stephen Warren <swarren@nvidia.com> Tested-by: Stephen Warren <swarren@nvidia.com>
sunxi: Machine id hack to prevent loading buggy sunxi-3.4 kernels
Right now U-Boot supports the CONFIG_OLD_SUNXI_KERNEL_COMPAT option,
which makes it go out of its way in limiting the selection of PLL clock
frequencies and PMIC voltages in order not to upset outdated buggy
sunxi-3.4 kernel releases. And if the CONFIG_OLD_SUNXI_KERNEL_COMPAT
option is not set, then booting such old kernels exhibits various
failures at runtime. This is very user unfriendly, and there were
already several incidents when people wasted their time being hit
by these runtime failures and trying to debug them.
The right solution is not to add hacks and workarounds to the mainline
U-Boot, but to fix these bugs in the sunxi-3.4 kernel. And in fact,
the updated sunxi-3.4 kernels already exist. Still we need to follow
the 'Principle of Least Surprise' and U-Boot needs to ensure that
the old buggy kernels are not getting happily booted when the
CONFIG_OLD_SUNXI_KERNEL_COMPAT option is not set. And this patch
addresses this particular issue.
This patch makes U-Boot store the 'compatibility revision' number in
the top 4 bits of the machine id and pass it to the kernel. The old
buggy kernels will fail to load with a very much googlable error
message on the serial console (the "r1 = 0x100010bb" part of it):
"Error: unrecognized/unsupported machine ID (r1 = 0x100010bb)"
This error message can be documented in the linux-sunxi wiki with
proper explanations about how to resolve this situation and where
to get the necessary bugfixes for the sunxi-3.4 kernel.
The fixed sunxi-3.4 kernels implement a revision compatibility check
and clear the top 4 bits of the machine id if everything is alright.
By accepting the machine id with the bits 31:28 set to 1, the sunxi-3.4
kernel effectively certifies that it has the PLL5 clock speed and
AXP209 DCDC3 voltage fixes applied.
It is still possible to set the CONFIG_OLD_SUNXI_KERNEL_COMPAT option
in U-Boot if the user desires to use an outdated unpatched sunxi-3.4
kernel.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com> Acked-by: Ian Campbell <ijc@hellion.org.uk> Acked-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Hans de Goede [Fri, 20 Feb 2015 15:55:12 +0000 (16:55 +0100)]
sunxi: Set the /chosen/stdout-path fdt property for sunxi boards
While discussing with some people how to get the Linux kernel to do the
right thing wrt sending output to both the serial console and the
hdmi out / lcd screen when booting on ARM devices, Grant Likely pointed out
that there already is a solution for this.
All we need to do is set the /chosen/stdout-path fdt property, and if no
console= arguments were specified on the kernel commandline the kernel
will honor this and add this device as a console (next to the primary
video output on hdmi).
And u-boot already has support for setting this, all we need to do is
define OF_STDOUT_PATH and then everything will just work ootb, without
people needing to meddle with adding console= arguments in extlinux.conf .
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Ian Campbell <ijc@hellion.org.uk> Reviewed-by: Tom Rini <trini@ti.com>
Hans de Goede [Thu, 19 Feb 2015 13:46:44 +0000 (14:46 +0100)]
sunxi: Fix sun5i mbus speed when booting old kernels
Older linux-sunxi-3.4 kernels override our PLL6 setting with 300 MHz,
halving the mbus frequency, so set it to 300 MHz ourselves and base the
mbus divider on that.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Ian Campbell <ijc@hellion.org.uk>
Hans de Goede [Mon, 16 Feb 2015 21:13:43 +0000 (22:13 +0100)]
sunxi: musb: Check Vbus-det before enabling otg port power
Sending out 5V when there is a charger connected to the otg port is not a
good idea, so check for this and error out.
Note this commit currently breaks otg support on the q8h tablets, as we need
to do some magic with the pmic there to get vbus info, this is deliberate
(better safe then sorry), fixing this is on my TODO list.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Ian Campbell <ijc@hellion.org.uk>
Hans de Goede [Thu, 19 Feb 2015 19:34:32 +0000 (20:34 +0100)]
sunxi: mmc: Always declare High Capacity capability
High Capacity (e)MMC cards work fine on sun4i / sun5i, and not having this
capability set causes u-boot to not recognize the eMMC on an Utoo P66 A13
tablet, so always set it thereby fixing this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Ian Campbell <ijc@hellion.org.uk>
Stephen Warren [Mon, 16 Feb 2015 19:16:15 +0000 (12:16 -0700)]
rpi: add support for Raspberry Pi 2 model B
USB doesn't seem to work yet; the controller detects the on-board Hub/
Ethernet device but can't read the descriptors from it. I haven't
investigated yet.
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Stephen Warren [Mon, 16 Feb 2015 19:16:14 +0000 (12:16 -0700)]
bcm2836 SoC support (used in Raspberry Pi 2 model B)
The bcm2835 and bcm2836 are essentially identical, except:
- The CPU is an ARM1176 v.s. a quad-core Cortex-A7.
- The physical address of many IO controllers has moved.
Rather than introducing a whole new bcm2836 value for $(SOC) or $(ARCH),
update the existing bcm2835 code to handle the minor differences, and
plumb it into the ARMv7 CPU architecture.
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Masahiro Yamada [Fri, 20 Feb 2015 08:04:19 +0000 (17:04 +0900)]
ARM: prepare for including <mach/*.h>
This commit adds $(srctree)/arch/arm/$(machdirs)/include/mach to
the headers search path.
It allows us to replace "#include <asm/arch/foo.h>" with
"#include <mach/foo.h>". As "#include <asm/arch/foo.h>" is still
supported, we can modify each file one by one.
Masahiro Yamada [Fri, 20 Feb 2015 08:04:12 +0000 (17:04 +0900)]
kbuild: prepare for moving headers into mach-*/include/mach
In U-Boot, SoC-specific headers are placed in
arch/$(ARCH)/include/asm/arch-$(SOC) and a symbolic link to that
directory is created at the early stage of the build process.
Creating and removing a symbolic link during the build is not
preferred. In fact, Linux Kernel did away with include/asm-$(ARCH)
directories a long time time ago.
As for ARM, now it is possible to collect SoC sources into
arch/arm/mach-$(SOC). It is also reasonable to move SoC headers
into arch/arm/mach-$(SOC)/include/mach.
This commit prepares for that.
If the directory arch/$(ARCH)/mach-$(SOC)/include/mach exists,
a symbolic to that directory is created. Otherwise, a symbolic link
to arch/$(ARCH)/include/asm/arch-$(SOC) or arch-$(CPU) is created.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Tested-by: Simon Glass <sjg@chromium.org> [ on nyan-big ] Cc: Stephen Warren <swarren@nvidia.com> Cc: Tom Warren <twarren@nvidia.com>
Masahiro Yamada [Fri, 20 Feb 2015 08:04:02 +0000 (17:04 +0900)]
ARM: prepare for moving SoC sources into mach-*
In U-boot, the directory structure, arch/$(ARCH)/cpu/$(CPU)/$(SOC)/
has been adopted except that $(CPU) is missing from some
architectures and $(SOC) is missing from some CPUs.
This structure did not fit very well in some cases.
[1] AT91
AT91 SoC family have been developed across some ARM processor
generations. Generally speaking, some IPs are often re-used in the
same SoC family (same SoC vendor) even when the main processor is
updated. As a result, a SoC-common directory is needed in the upper
level. Currently, AT91 source files are placed as follows:
Once directories are split, the motivation for refactorings across
CPU directories is lost. Some files in arm920t/at91/ and
arm926ejs/at91/ are so similar that they could be merged.
[2] Tegra
Tegra is a little bit special case where different CPUs are used for
SPL and the main U-boot. To obey the arch/$(ARCH)/cpu/$(CPU)/$(SOC)
structure, the source files must be placed across the CPUs,
again SoC-common directory is necessary in the upper level.
Moreover, there are several families in Tegra: Tegra20, Tegra30,
Tegra114, Tegra124. Here again, the tegra-common directory is needed
to contain commonly-used files.
Tegra directories have been sprinkled in the directory structure.
Masahiro Yamada [Fri, 20 Feb 2015 08:04:01 +0000 (17:04 +0900)]
ARM: at91: move board select menu and common settings
The board select menu in arch/arm/Kconfig is still big.
To slim down it, this commit moves AT91 boards to
arch/arm/mach-at91/Kconfig.
Also, consolidate "config SYS_SOC" in each board Kconfig.
The Kconfig files under board/ directory were modified with the
following command:
find board -name Kconfig | xargs sed -i -e '
/config SYS_SOC/ {
N
/default "at91"/ {
N
d
}
}
'
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Andreas Bießmann <andreas.devel@googlemail.co>
Marek Vasut [Wed, 18 Feb 2015 21:36:18 +0000 (22:36 +0100)]
dm: Protect device_unbind() with CONFIG_DM_DEVICE_REMOVE
Since device_unbind() is also defined in device-remove.c,
which is compiled in only in case CONFIG_DM_DEVICE_REMOVE
is defined, protect the device_unbind() prototype with the
same CONFIG_DM_DEVICE_REMOVE check.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Stefan Roese <sr@denx.de> Cc: Tom Rini <trini@ti.com> Acked-by: Simon Glass <sjg@chromium.org>
Simon Glass [Tue, 27 Jan 2015 01:27:09 +0000 (18:27 -0700)]
serial: ns16550: Support debug UART
Add debug UART functions to permit ns16550 to provide an early debug UART.
Try to avoid using the stack so that this can be called from assembler before
a stack is set up (at least on ARM and PowerPC).
Simon Glass [Tue, 27 Jan 2015 01:27:08 +0000 (18:27 -0700)]
serial: ns16550: Add access functions that don't need platdata
For the debug UART we need to be able to provide any parameters before
driver model is set up. Add parameters to the low-level access functions
to make this possible.
Simon Glass [Tue, 27 Jan 2015 01:27:07 +0000 (18:27 -0700)]
serial: Support an early UART for debugging
This came up in a discussion on the mailing list here:
https://patchwork.ozlabs.org/patch/384613/
My concerns at the time were:
- it doesn't need to be written in assembler
- it doesn't need to be ARM-specific
This patch provides a possible alternative. It works by allowing any serial
driver to export one init function and provide a putc() function. These
can be used to output debug data before the real serial driver is available.
This implementation does not depend on driver model, and it is possible for
it to operate without a stack on some architectures (e.g. PowerPC, ARM). It
provides the same features as the ARM-specific debug.S but with more UART
and architecture support.
Simon Glass [Fri, 13 Feb 2015 19:20:48 +0000 (12:20 -0700)]
dm: Move CONFIG_I2C_COMPAT to Kconfig
Make this option available in Kconfig and clean up the board that uses it.
Note there is also an entry in exynos5-common.h but this affects multiple
boards and should be dropped as part of the Samsung I2C migration to
driver model.
Lubomir Popov [Fri, 30 Jan 2015 17:56:04 +0000 (19:56 +0200)]
cmd_i2c: Provide option for bulk 'i2c write' in one transaction
I2C chips do exist that require a write of some multi-byte data to occur in
a single bus transaction (aka atomic transfer), otherwise either the write
does not come into effect at all, or normal operation of internal circuitry
cannot be guaranteed. The current implementation of the 'i2c write' command
(transfer of multiple bytes from a memory buffer) in fact performs a separate
transaction for each byte to be written and thus cannot support such types of
I2C slave devices.
This patch provides an alternative by allowing 'i2c write' to execute the
write transfer of the given number of bytes in a single bus transaction if
the '-s' option is specified as a final command argument. Else the current
re-addressing method is used.
Signed-off-by: Lubomir Popov <l-popov@ti.com>
hs: adapt to CONFIG_DM_I2C
Joe Hershberger [Thu, 5 Feb 2015 03:56:54 +0000 (21:56 -0600)]
cmd_fdt: Print the control fdt in terms of virtual memory
If you want to inspect the control device tree using the fdt command,
the "fdt address -c" command previously unhelpfully printed the phys
memory address of the device tree. That address could not then be used
to set the fdt address for inspection. Changed the resulting print to
one that can be copied directly to the 'fdt address <addr>' command.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com> Acked-by: Simon Glass <sjg@chromium.org>
Before avr32 had an extra storage for stack end to have a nice stack printout
on exception. Remove this extra storage and use generic gd->start_addr_sp
instead.
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
cpu_mmc_init() is required by the init sequence to have a working MMC interface
on avr32. This will not be included in the binary if we omit the avr32 board.c
when building the generic board.
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
For example on a raspberry pi the u-boot environment can be
saved in a file on the first VFAT partition.
This example illustrates how to use it with fw_printenv/fw_setenv.
projects / u-boot / log