We run 4 Arndale boards in our automated test framework, they have
been running quite happily for quite some time using a Debian Wheezy
userspace.
However when upgrading to a Debian Jessie we started seeing frequent
segmentation faults from gcc when building the kernel, to the extent
that it is unable to successfully build the kernel twice in a row, and
often fails on the first attempt.
Searching around I found https://bugs.launchpad.net/arndale/+bug/1081417
which pointed towards http://www.spinics.net/lists/kvm-arm/msg03723.html
and CPU Errata 773022 and 774769.
This errata needs to be applied to all processors in an SMP system,
meaning that the usual strategy of applying them in
arch/arm/cpu/armv7/start.S is not appropriate (since that applies to
the boot processor only). Instead we apply these errata in the secure
monitor which is code that is traversed by all processors as they are
brought up.
The net affect on Arndale is that ACTLR changes from 0x40 to
0x2000042. I ran 17 kernel compile iterations overnight with no
segfaults.
Runtime testing was done on our v2014.10 based branch and forward
ported (with only minimal and trivial contextual conflicts) to current
master, where it has been build tested only.
I suppose in theory these errata apply to any Exynos5250 based boards,
but Arndale is the only one I have access to and I have therefore
chosen to be conservative and only apply it there.
Also, reorder CONFIG_ARM_ERRATA_794072 in README to make the list
numerically sorted.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
If an Android boot image does not contain a ramdisk, make sure rd_len
and rd_data are returned to indicate no ramdisk rather than just relying
on returning an error.
Signed-off-by: Rob Herring <robh@kernel.org>
This patch adds support for LZ4-compressed FIT image contents. This
algorithm has a slightly worse compression ration than LZO while being
nearly twice as fast to decompress. When loading images from a fast
storage medium this usually results in a boot time win.
Sandbox-tested only since I don't have a U-Boot development system set
up right now. The code was imported unchanged from coreboot where it's
proven to work, though. I'm mostly interested in getting this recognized
by mkImage for use in a downstream project.
Signed-off-by: Julius Werner <jwerner@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
This platform has not gone into production. So lets remove it.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Add CONFIG_SYS_GENERIC_BOARD to lwmon5.h and CONFIG_DISPLAY_BOARDINFO
to Kconfig file.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
This reverts commit 8fe11b8901.
I'll add support to lwmon5 in the next patch and will remove
support for the broken lcd4_lwmon5 as well.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Overwriting an empty file not created by U-Boot did not work, and it
could even corrupt the FAT. Moreover, creating empty files or emptying
existing files allocated a cluster, which is not standard.
Fix this by always keeping empty files clusterless as specified by
Microsoft (the start cluster must be set to 0 in the directory entry in
that case), and by supporting overwriting such files.
Signed-off-by: Benoît Thébaudeau <benoit@wsystem.com>
curclust was used instead of newclust in the debug() calls and in one
CHECK_CLUST() call, which could skip a failure case.
Signed-off-by: Benoît Thébaudeau <benoit@wsystem.com>
set_contents() had uselessly split calls to set_cluster(). Merge these
calls, which removes some cases of set_cluster() being called with a
size of zero.
Signed-off-by: Benoît Thébaudeau <benoit@wsystem.com>
set_cluster() was using a temporary buffer without enforcing its
alignment for DMA and cache. Moreover, it did not check the alignment of
the passed buffer, which can come directly from applicative code or from
the user.
This could cause random data corruption, which has been observed on
i.MX25 writing to an SD card.
Fix this by only passing ARCH_DMA_MINALIGN-aligned buffers to
disk_write(), which requires the introduction of a buffer bouncing
mechanism for the misaligned buffers passed to set_cluster().
By the way, improve the handling of the corresponding return values from
disk_write():
- print them with debug() in case of error,
- consider that there is an error is disk_write() returns a smaller
block count than the requested one, not only if its return value is
negative.
After this change, set_cluster() and get_cluster() are almost
symmetrical.
Signed-off-by: Benoît Thébaudeau <benoit@wsystem.com>
CONFIG_TWL4030_POWER is a boolean define variable. It is either defined
or not defined and should not have a value assigned to it.
Remove the value.
Signed-off-by: Igor Grinberg <grinberg@compulab.co.il>
This patch changes the default "root=" parameter to "/dev/sda2".
Many linux based distros use /dev/sda1 for their boot partition; this is
often not a rootfs that can be used by the "root=" parameter.
Linaro images use /dev/sda1 as a boot partition, although this of a
different nature to a distro image. Linaro uses /dev/sda2 for the rootfs
partition.
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
The latest Juno firmware stores the files in NOR flash as "norkern" for
kernel binary, "board.dtb" for the device tree binary.
The "old" firmware used the name "Image" for the kernel binary and
"juno" for the device tree binary.
Rather than just change the default U-Boot configuration to use the new
names, breaking users with the old firmware, attempt to load the default
filename first. If that fails, attempt to load the alternate filename.
I've echo'd that we are loading the alternate file to counter the
output from "afs load" shown if the first load attempt fails. For
example, I see output like this on my Juno board when it's configured
the with the "old" firmware:
image "norkern" not found in flash
Loading Image instead of norkern
loaded region 0 from 08500000 to 80000000, 00AB6318 bytes
image "board.dtb" not found in flash
Loading juno instead of board.dtb
loaded region 0 from 0A000000 to 83000000, 00003188 bytes
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Some OS images require an initrd on Juno.
If the file ramdisk.img exists in NOR flash, then we load it and pass
the address to the kernel. Otherwise, we pass the "-" parameter as
before.
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Change the load_image so that it returns success or failure of the
command (using CMD_RET_SUCCESS or CMD_RET_FAILURE).
This way, hush scripts can optionally load different files depending
upon the system configuration.
A simple example:
if afs load ${kernel_name} ${kernel_addr}; then echo loaded; else echo \
not loaded; fi
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Add a command to the ARM flash support to check if an image exists or
not.
If the image is found, it will return CMD_RET_SUCCESS, else
CMD_RET_FAILURE. This allows hush scripts to conditionally load images.
A simple example:
if afs exists ${kernel_name}; then echo found; else echo \
not found; fi
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Linaro's Juno Android builds requires the androidboot.hardware parameter
be set to a know board name.
Non-Android kernels ignore this extra parameter because they don't
contain code to parse it.
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Create an additional FVP configuration to boot images pre-loaded into
DRAM.
Sometimes it's preferential to boot the model by loading the files
directly into DRAM via model parameters, rather than using
SemiHosting.
An example of model parmaters that are used to pre-load the files
into DRAM:
--data cluster0.cpu0=Image@0x80080000 \
--data cluster0.cpu0=fvp-base-gicv2-psci.dtb@0x83000000 \
--data cluster0.cpu0=uInitrd@0x84000000
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
[trini: Update board/armltd/vexpress64/Kconfig logic]
Signed-off-by: Tom Rini <trini@konsulko.com>
vexpress64 kernels are usually over 8 MBytes in length, so setting the
max uImage length to 64 Mbytes should give us plenty of scope for
expansion.
I mostly chose this length to match other board configs that use
"(64 << 20)".
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
The FVP and Juno settings were identical, but duplicated, so I removed
the duplication with this patch.
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
[trini: Adjust logic to keep if/endif in the file]
Signed-off-by: Tom Rini <trini@konsulko.com>
This patch fixes a couple of checkpatch warnings on the vexpress64 config.
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
AT26DF081A is the spi flash type of TWR-MEM(SCH-26248) card.
We can access the flash through DSPI2 on LS1021ATWR board.
Signed-off-by: Yuan Yao <yao.yuan@freescale.com>
Reviewed-by: Jagan Teki <jteki@openedev.com>
Erratum A-008022 has been fixed on LS1021A Rev2.0.
So we can use DSPI2 now, this patch enable DSPI2
in dts for LS1021ATWR.
Signed-off-by: Yuan Yao <yao.yuan@freescale.com>
Reviewed-by: Jagan Teki <jteki@openedev.com>
priv->mode is initialized when .set_speed triggers
with mode value, so checking mode for configuring
CPOL, CPHA using priv->mode is invalid hence use
mode from .set_speed argument, and at the end
priv->mode will initialized with mode.
This patch also replaces formatting string to use
speed instead of mode in .set_speed ops.
Signed-off-by: Jagan Teki <jteki@openedev.com>
priv->mode is initialized when .set_speed triggers
with mode value, so checking mode for configuring
CPOL, CPHA using priv->mode is invalid hence use
mode from .set_speed argument, and at the end
priv->mode will initialized with mode.
This patch also replaces formatting string to use
speed instead of mode in .set_speed ops.
Signed-off-by: Jagan Teki <jteki@openedev.com>
The pcDuino1 board unconditionally provides 5V to USB host
receptacles. The pcDuino2 board has a voltage regulator,
controlled by the PD2 pin which is pulled-up by default
(so that the USB power is also enabled by default).
Not specifying pins for enabling USB power in the defconfig
means that the PH3 and PH6 pins are driven high by default.
The PH6 pin is available on the Arduino-compatible expansion
header and touching it is not nice (this may be even dangerous,
depending on what kind of role is assigned to this particular
pin by various Arduino shields).
This patch explicitly configures the USB VBUS pins to "",
which means that no pins should be touched. The patch has
been tested on a pcDuino2 board and USB still works.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
PCI_HEADER_TYPE register (offset 0x0e) bit 7 is an indicator
for multi-function devices. We should mask it off before using
it as the header type.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>
It turned out with some boards (FPGA firmwares?) and cards combos
current clock settings doesn't work as expected leading to strange
card freezes or corrupted data being read from the card.
Especially this was seen with Transcend 2Gb cards shipped as a part of
ARC SDP:
----------------->8---------------
AXS# mmcinfo
Device: Synopsys Mobile storage
Manufacturer ID: 74
OEM: 4a60
Name: SDC
Tran Speed: 50000000
Rd Block Len: 512
SD version 3.0
High Capacity: No
Capacity: 1.8 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
AXS# fatload mmc 0
** Unrecognized filesystem type **
----------------->8---------------
With this change that problem is fixed.
Note "Tran Speed" above doesn't match clock value set in DW MMC.
It is max value for card's speed class.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
When CHECK_BITS_SET was added, they forgot to add
a new command table, and instead overwrote the
previous table.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
Tested-by: Fabio Estevam <fabio.estevam@freescale.com>
Currently 'reset' only works with the test device tree. When run without a
device tree, or with the normal device tree, the following error is
displayed:
Reset not supported on this platform
Fix the driver and the standard device tree to avoid this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reported-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@wwwdotorg.org>
Adjust the memory leak tests to show the amount of memory leaked. This can
be a useful signal as to what is wrong.
Signed-off-by: Simon Glass <sjg@chromium.org>
Currently when driver model starts up it finds the root uclass and the
pinctrl uclass. This is because even the root node handles pinctrl
processing.
But this is not useful. The root node is not a real hardware device so
cannot require any particular pinmux settings. Also it means that the
memory leak tests fails, since they end up freeing more memory than
they allocate: the marker it set after the root device and pinctrl
uclass are allocated, and later once the pinctrl uclass is freed the memory
used by driver model is less than when the marker was set.
If a platform needs 'core' pin mulitplex settings it can do this with
a driver that is probed on start-up. It would be an abuse of the root node
to use this for pinctrl.
To avoid this problem, only process pinctrl settings for non-root nodes.
Signed-off-by: Simon Glass <sjg@chromium.org>
When malloc_base initially gets setup in the SPL it is based on the
current (early) stack pointer, which for rockchip is pointing into SRAM.
This means simple memory allocations happen in SRAM space, which is
somewhat unfortunate. Specifically a bounce buffer for the mmc allocated
in SRAM space seems to cause the mmc engine to stall/fail causing
timeouts and a failure to load the main u-boot image.
To resolve this, reconfigure the malloc_base to start at the relocated
stack pointer after DRAM has been setup.
For reference, things did work fine on rockchip before 596380db was
merged to fix memalign_simple due to a combination of rockchip SDRAM
starting at address 0 and the dw_mmc driver not checking errors from
bounce_buffer_start. As a result, when a bounce buffer needed to be
allocated mem_align simple would fail and return NULL. The mmc driver
ignored the error and happily continued with the bounce buffer address
being set to 0, which just happened to work fine..
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Simon Glass <sjg@chromium.org>
It looks like this line was copy-pasted, but not modified.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Simon Glass <sjg@chromium.org>
This device uses SDHCI driver, for eMMC and SD cards.
Trying bind the DW MMC driver with fdt node without all
required properties, causes printing an error.
This commit disables the DW MMC node.
Tested-on: Trats
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Łukasz Majewski <l.majewski@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
After rework of code by:
commit: d952796 Exynos5: Use clock_get_periph_rate generic API
function get_mmc_clk() always returns -1 for Exynos 4.
This was caused by omitting, that SDHCI driver for Exynos 4,
calls get_mmc_clk(), with mmc device number as argument,
instead of pinmux peripheral id, like DW MMC driver for Exynos 5.
By this commit, the code directly calls a proper function
to get mmc clock for Exynos 4, without checking the peripheral id.
Tested on: Odroid U3/X2, Trats, Trats2, Odroid XU3, Snow (by Simon).
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
After rework in lib/fdtdec.c, the function fdtdec_get_addr()
doesn't work for nodes with #size-cells property set to 0.
To get GPIO's 'reg' property, the code should use one of:
fdtdec_get_addr_size_auto_no/parent() function.
Fortunately dm core provides a function to get the property.
This commit reworks function gpio_exynos_bind(), to properly
use dev_get_addr() for GPIO device.
This prevents setting a wrong base register for Exynos GPIOs.
Tested on: Odroid U3/X2, Trats, Trats2, Odroid XU3, Snow (by Simon).
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
After rework of lib/fdtdec.c by:
commit: 02464e3 fdt: add new fdt address parsing functions
the function fdtdec_get_addr() doesn't work as previous,
because the implementation assumes that properties '#address-cells'
and '#size-cells' are equal to 1, which can be not true sometimes.
The new API introduced fdtdec_get_addr_size_auto_parent() for the 'reg'
property parsing, but the implementation assumes, that #size-cells
can't be less than 1.
This causes that the following children's 'reg' property can't be reached:
parent@0x0 {
#address-cells = <1>;
#size-cells = <0>;
children@0x100 {
reg = < 0x100 >;
};
};
Change the condition value from '1' to '0', which allows parsing property
with at least zero #size-cells, fixes the issue.
Now, fdtdec_get_addr_size_auto_parent() works properly.
Tested on: Odroid U3/X2, Trats, Trats2, Odroid XU3, Snow (by Simon).
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
fdtdec_get_addr_size() may be used in two cases:
a) With sizep supplied, in which case both an address and a size are
parsed from DT. In this case, the DT property must be large enough to
contain both values.
b) With sizep NULL, in which case only an address is parsed from DT.
In this case, the DT property only need be large enough to contain this
address value. Commit 02464e386b "fdt: add new fdt address parsing
functions" broke this relaxed checking, and required the DT property to
contain both an address and a size value in all cases.
Fix fdtdec_get_addr_size() to vary ns based on whether the size value
is being parsed from the DT or not. This is safe since the function only
parses the first entry in the property, so the overall value of (na + ns)
need not be accurate, since it is never used to step through the property
data to find other entries. Besides, this fixed behaviour essentially
matches the original behaviour before the patch this patch fixes. (The
original code validated that the property was exactly the length of
either na or (na + ns), whereas the current code only validates that the
property is at least that long. For non-failure cases, the two behaviours
are identical).
Cc: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Thierry Reding <treding@nvidia.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Michal Suchanek <hramrach@gmail.com>
Fixes: 02464e386b ("fdt: add new fdt address parsing functions")
Reported-by: Przemyslaw Marczak <p.marczak@samsung.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Przemyslaw Marczak <p.marczak@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
This comment from README.fdt-control did not end up in the Kconfig, which
is what most people will see. Add it with a few tweaks.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Michal Simek <michal.simek@xilinx.com>
The 7" Q8 tablet enclosure is used for a ton of slightly different cheap
chinese tablets. There are some differences in which accelerometer /
wifi is used, but other then that these are all the same from a u-boot /
kernel pov.
When we get to adding accelerometer support the plan is to add some kind
of autodetection and mangle the dt accordingly (likely using the new quirks
mechanism).
For now this is a non issue as we do not yet have accelerometer
support, and in the future, some sort of auto-detect is the way to go
as we cannot expect users to exactly know what is inside their tablet.
The dts files this commit adds are identical to the ones submitted
to the upstream kernel.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
