xilinx: zynqmp: Add new target with only emmc enabled
This patch adds new target which is called as mini configuration
with only emmc functionality and other required basic features enabled.
This will be used to run in system with small footprint and needs
emmc support.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
xilinx: zynqmp: Add new target with only nand enabled
This patch adds new target which is called as mini configuration
with only nand functionality and other required basic features enabled.
This will be used to run in system with small footprint and needs
nand support.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
arm64: zynqmp: Provide a config to not map DDR region in MMU table
DDR less systems are possible for configuration like mini qspi
and making DDR region as normal memory may cause speculative
access which results u-boot hang if DDR is absent. So, this
patch fixes the issue by not making DDR memory region
entry into MMU table.
Future solution is to prepare MMU table per memory node in dts instead
of hard code DDR addresses.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Tue, 9 Jan 2018 14:27:31 +0000 (15:27 +0100)]
arm: zynq: Enable DM_GPIO when needed
There are two reasons for doing this change.
There is still !DM driver for xilinx soft gpio IP and especially it is
saving some space for memory constrained boards like cse (almost ~400B).
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Tue, 9 Jan 2018 13:52:29 +0000 (14:52 +0100)]
arm: zynq: Enable BLK when needed
There is no reason to enable BLK by default for all boards which is just
increasing memory footprint for memory contrained boards like cse.
zc770s are also saving some space.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
xilinx: zynqmp: Use strlen only if env_get doesn't return null
Add check if boot_targets exists in environment and then
generate new_targets env accordingly. Performing strlen on
null address causes it to fail with exception if isolation
is enabled with DDR address zero as secure. It works with out
isolation enabled as zero is valid address but it may lead to
junk values in boot_targets.
This patch fixes the issue by checking return value of env_get
so that it generate boot_targets properly.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Fri, 12 Jan 2018 13:24:50 +0000 (14:24 +0100)]
arm: zynq: Fix types in ps7_spl_init
The patch is fixing the following Warning:
arch/arm/mach-zynq/ps7_spl_init.c:133:24: warning: comparison between
signed and unsigned integer expressions [-Wsign-compare]
while (ioread(addr) < delay)
^
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Andy Shevchenko [Wed, 10 Jan 2018 17:40:15 +0000 (19:40 +0200)]
x86: zImage: Propagate acpi_rsdp_addr to kernel via boot parameters
New field acpi_rsdp_addr, which has been introduced in boot protocol
v2.14 [1], in boot parameters tells kernel the exact address of RDSP
ACPI table. Knowing it increases robustness of the kernel by avoiding
in some cases traversal through a part of physical memory.
It will slightly reduce boot time by the same reason.
[1] See Linux kernel commit
2f74cbf ("x86/boot: Add the ACPI RSDP address to struct setup_header::acpi_rdsp_addr")
@ https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?id=2f74cbf
for the details.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: updated the kernel commit git URL and fixed one style issue] Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
introduced an assignment of subarch field in boot parameters, though
missed the right place of doing that. It doesn't matter if we have or
not a kernel command line supplied, we just set that field. Although
guard it by protocol version which supports it.
Fixes: 20bfac0599bd ("x86: zImage: add Intel MID platforms support") Cc: Vincent Tinelli <vincent.tinelli@intel.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Andy Shevchenko [Wed, 10 Jan 2018 16:07:26 +0000 (18:07 +0200)]
x86: tangier: Make _CRS for BTH0 Serialized to avoid warning
ASL compiler warns:
ASL board/intel/edison/dsdt.asl
board/intel/edison/dsdt.asl.tmp 238: Method (_CRS, 0, NotSerialized)
Remark 2120 - Control Method should be made Serialized ^ (due to creation of named objects within)
Do as suggested by ASL compiler.
Fixes: 5d8c4ebd95e2 ("x86: tangier: Add Bluetooth to ACPI table") Reported-by: Ferry Toth <fntoth@gmail.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Андрей Мозжухин [Wed, 3 Jan 2018 12:43:56 +0000 (15:43 +0300)]
aes: Allow non-zero initialization vector
AES encryption in CBC mode, in most cases, must be used with random
initialization vector. Using the same key and initialization vector several
times is weak and must be avoided.
Added iv parameter to the aes_cbc_encrypt_blocks and aes_cbc_decrypt_blocks
functions for passing initialization vector.
Command 'aes' now also require the initialization vector parameter.
Masahiro Yamada [Sun, 28 Jan 2018 09:41:23 +0000 (18:41 +0900)]
kconfig: revert change that was not needed for -Wformat-security
Recent GCC versions warn if the format string is not a literal
because the compiler cannot check the argument validity at compile
time.
Commit 192bc6948b02 ("Fix GCC format-security errors and convert
sprintfs.") blindly replaced sprintf() with strcpy(), including
many cases where the format parameter is a string literal.
For the kconfig change:
sprintf(header, " ");
..., here the format parameter is a string literal " ", so it is
definitely equivalent to:
strcpy(header, " ");
Of course, if the 'header' did not have enough length for containing
" ", it would be a security problem, but another problem. (in this
case, the 'header' is 4 byte length buffer, so it is not a problem at
all.)
The kconfig code is kept as synced with Linux as possible, but this
change made the code out-of-sync for nothing. Just reverting.
Tom Rini [Sat, 27 Jan 2018 19:48:08 +0000 (14:48 -0500)]
mvpp2: Fix warning over 32bit vs 64bit targets
When we have a driver that is used on both 32bit and 64bit targets and
we are talking about address space we cannot use u64 nor u32 and instead
need to use phys_addr_t.
Fixes: 377883f16d36 ("net: mvpp2x: fix phy connected to wrong mdio issue") Cc: Stefan Chulski <stefanc@marvell.com> Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Stefan Roese <sr@denx.de> Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Patrice Chotard [Mon, 29 Jan 2018 17:14:14 +0000 (18:14 +0100)]
clk: clk_stm32: Add .set_rate callback
Since 'commit f4fcba5c5baa ("clk: implement clk_set_defaults()")'
STM32F4 family board can't boot.
Above patch calls clk_set_rate() for all nodes with assigned-clock-rates
property. Clock driver for STM32F family doesn't implement .set_rate
callback which make clk_set_defaults() exit on error and prevent board
to boot.
Patrice Chotard [Mon, 29 Jan 2018 12:44:20 +0000 (13:44 +0100)]
board: stm32: Fix stm32f746-disco boot
Since 'commit af2f44267 ("fdc spl: use different BOARD_INIT MACRO for spl and tpl")'
board stm32f746-disco can't boot.
The macro CONFIG_IS_ENABLED() can't evaluate the value of
CONFIG_SPL_BOARD_INIT as it was defined in include/configs/stm32f746-disco.h
without attributed value.
Moving CONFIG_SPL_BOARD_INIT from stm32f746-disco.h to mach-stm32/Kconfig
fixes this issue.
Fixes: af2f44267 ("fdc spl: use different BOARD_INIT MACRO for spl and tpl") Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Alexander Graf [Mon, 29 Jan 2018 12:57:20 +0000 (13:57 +0100)]
serial: Make full device search optional
Commit 608b0c4ad4e5ec0c ("serial: Use next serial device if probing fails")
added code to search for more serial devices if the default one was not
probed correctly.
Unfortunately, that breaks omap3_evm. So while investigating why that is
the case, let's disable the full search for everyone but bcm283x where it
is needed.
Fixes: 608b0c4ad4e5ec0c ("serial: Use next serial device if probing fails") Reported-by: Derald D. Woods <woods.technical@gmail.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Andreas Färber [Sat, 20 Jan 2018 12:07:09 +0000 (13:07 +0100)]
tools/mrvl_uart.sh: Fix minicom baudrate
minicom doesn't inherit the baudrate from stty but uses its own
defaults, such as for example 57600, whereas we expect 115200 here.
Explicitly tell minicom which baudrate to use.
Fixes: eee4835d22 ("tools: Add Marvell recovery image download script") Cc: Konstantin Porotchkin <kostap@marvell.com> Cc: Stefan Roese <sr@denx.de> Cc: Igal Liberman <igall@marvell.com> Signed-off-by: Andreas Färber <afaerber@suse.de> Reviewed-by: Konstantin Porotchkin <kostap@marvell.com> Reviewed-by: Stefan Roese <sr@denx.de> Signed-off-by: Stefan Roese <sr@denx.de>
The variable cfi_flash_num_flash_banks is defined iff
CONFIG_SYS_MAX_FLASH_BANKS_DETECT is defined, but it is used
unconditionally in the function cfi_flash_init_dm. This leads to a
undefined variable compile error when CONFIG_SYS_MAX_FLASH_BANKS_DETECT
is not defined, but DM is enabled.
Fix this by always defining the cfi_flash_num_flash_banks variable.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:55 +0000 (14:43 +0100)]
cfi_flash: Fix indention
When long expressions surrounded by parentheses are split into multiple
lines, each consecutive line should be aligned with the corresponding
parenthesis. Fix all instances where this occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:53 +0000 (14:43 +0100)]
cfi_flash: Bound-check index before array access
In a while loop in cfi_flash.c the array "start" is accessed at the index
"sector" before the index variable "sector" is bounds-checked, which
might lead to accesses beyond the bounds of the array.
Swap the order of the checks in the "&&" expression, so that the
short-circuit evaluation prevents out-of-bounds array accesses.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:49 +0000 (14:43 +0100)]
cfi_flash: Use u8 pointers instead of void pointers
According to the C standard, pointer arithmetic for pointers of type
void is undefined behavior (the assumption that they're 8-bit wide is a
GCC-specific assumption). In the interest of keeping the code
standards-compliant, and also better communicate intent, switch all
void* variables where pointer arithmetic is used to u8* variables.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:48 +0000 (14:43 +0100)]
cfi_flash: Remove assignments from if conditions
The condition in if statements should not be used for variable
assignment. Instead, the assignment should be done in a separate step
beforehand. Fix all instances where this occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:45 +0000 (14:43 +0100)]
cfi_flash: Fix else after break
If in a loop, the if block in a if/else statement ends in a break, the
statements in the else blockcan be extracted, since the break stops the
execution.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:43 +0000 (14:43 +0100)]
cfi_flash: Add missing braces in blocks
In if/else statements, either both blocks (if and else) should have
braces or both blocks should not have braces, but mixed configurations
are discouraged. Fix all instances where this occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:42 +0000 (14:43 +0100)]
cfi_flash: Remove unnecessary braces
"==" and "!=" bind tighter than the boolean operators, so parentheses
around them in compound logical statements are unnecessary. Fix all
instances where this occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:41 +0000 (14:43 +0100)]
cfi_flash: Fix comment style
Comment blocks should end with a "*/" on a separate line, not with the
"*/" attached to the end of the last line of text. Fix all instances
where this occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:40 +0000 (14:43 +0100)]
cfi_flash: Use __func__ macro instead of function name
printf/debug statements should not include the file name as a hardcoded
string, but instead use the __func__ macro. Fix all instances where this
occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:39 +0000 (14:43 +0100)]
cfi_flash: Fix logical continuations
When splitting long logical statements across multiple lines, the
logical operators should be at the end of the lines. Fix all instances
where this occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
Mario Six [Fri, 26 Jan 2018 13:43:37 +0000 (14:43 +0100)]
cfi_flash: Fix missing/superfluous lines
There should be no consecutive blank lines, and no blank lines at the
end of blocks. But there should be blank lines between variable
declarations and code. Fix all instances where either occurs.
Signed-off-by: Mario Six <mario.six@gdsys.cc> Signed-off-by: Stefan Roese <sr@denx.de>
vsprintf.c: correct printing of a NULL device path
When printing '%pD' with a value of NULL we want to output
'<NULL>'. But this requires copying to buf. Leave this
to string16.
A unit test is supplied which relies on EFI support in the sandbox.
The development for EFI support in the sandbox is currently in branch
u-boot-dm/efi-working. The branch lacks commit 6ea8b580f06b ("efi_loader:
correct DeviceNodeToText for media types"). Ater rebasing the aforementioned
branch on U-Boot v2018.01 and adding 256060e4257a2 and this patch the test
is executed successfully.
Fixes: 256060e4257a2 (vsprintf.c: add EFI device path printing) Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Thu, 25 Jan 2018 23:47:53 +0000 (00:47 +0100)]
efi_loader: Call Exit() on return from payload in StartImage()
When a UEFI payload just returns instead of calling the Exit() callback,
we handle that in efi_do_enter() and call Exit on its behalf, so that
the loaded_image->exit_status value is correct.
We were missing that logic in StartImage(). Call it there too.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Alexander Graf <agraf@suse.de>
If 'bootefi hello' or 'bootefi selftest' can be executed depends on the
configuration.
If an invalid non-numeric 1st argument is passed to bootefi, e.g.
'bootefi hola', this string is converted to 0 and U-Boot jumps to
this typically invalid address.
With the patch the online help is shown instead.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Tue, 23 Jan 2018 23:18:08 +0000 (00:18 +0100)]
efi_loader: Always use EFIAPI instead of asmlinkage
EFI calls are usually defined as asmlinkage. That means we pass all parameters
to functions via the stack x86_32.
On x86_64 however, we need to also stick to the MS ABI calling conventions,
which the EFIAPI define conveniently handles for us. Most EFI functions were
also marked with EFIAPI, except for the entry call.
So this patch adjusts all entry calls to use EFIAPI instead of the manual
asmlinkage attribute.
While at it, we also change the prototype of the entry point to return
efi_status_t instead of ulong, as this is the correct prototype definition.
Alexander Graf [Wed, 24 Jan 2018 13:54:21 +0000 (14:54 +0100)]
efi: Conflict efi_loader with different stub bitness
We have 2 users of the EFI headers: efi_loader and the EFI stub. Efi_loader
always expects that the bitness of the definitions it uses is identical to
the execution.
The EFI stub however allows to run x86_64 U-Boot on 32bit EFI and the other
way around, so it allows for different bitness of EFI definitions and U-Boot
environment.
This patch explicitly requests via Kconfig that efi_loader can only be enabled
if the bitness is identical. Because we can run efi_loader on x86_64 without
EFI stub enabled, it also ensures that this case propagates the correct ABI
constraints.
Alexander Graf [Thu, 25 Jan 2018 11:05:56 +0000 (12:05 +0100)]
bcm2835_pinctrl: Probe pre-reloc
The serial drivers now depend on the pinctrl driver to determine whether
they are enabled. That means if a serial device wants to be used pre-reloc,
we also need the pinctrl device pre-reloc.
Adapt the pinctrl driver as well as dts overlay accordingly.
Alexander Graf [Thu, 25 Jan 2018 11:05:55 +0000 (12:05 +0100)]
bcm2835_pl011_serial: Add BCM2835 specific serial driver
On bcm2835 we need to ensure we only access serial devices that are
muxed to the serial output pins of the pin header. To achieve this
for the pl011 device, add a bcm2835 specific pl011 wrapper device
that does this check but otherwise behaves like a pl011 device.
Alexander Graf [Thu, 25 Jan 2018 11:05:54 +0000 (12:05 +0100)]
MAINTAINERS: Take over BCM2835 maintainership
It seems as if I have more interest in BCM2835 support than most others,
so I'll bite the bullet and declare myself maintainer. It'd be a shame
to leave that platform orphaned.
Alexander Graf [Thu, 25 Jan 2018 11:05:47 +0000 (12:05 +0100)]
serial_bcm283x_mu: Always skip init
The serial initialization doesn't always quite work for me, so let's
always skip it for now. We know that firmware on the RPi initializes
us properly already.
Alexander Graf [Thu, 25 Jan 2018 11:05:45 +0000 (12:05 +0100)]
rpi: Determine PL011/Mini-UART availability at runtime
Firmware on the Raspberry Pi family of devices can dynamically configure either
the PL011, Mini-UART or no device at all to be routed to the user accessible
UART pins.
That means we need to always include both drivers, because we can never be sure
which of the two serial devices firmware actually chooses to use.
Alexander Graf [Thu, 25 Jan 2018 11:05:44 +0000 (12:05 +0100)]
serial: bcm283x_mu: Remove support for post-init disabling
We are switching to a model where a serial device doesn't even get probed when
it's not muxed properly, so we don't need device specific disabling
functionality anymore.
Alexander Graf [Thu, 25 Jan 2018 11:05:43 +0000 (12:05 +0100)]
rpi: Remove runtime disabling support for serial
We are switching to a model where our board file can directly fail probing
of serial devices when they're not usable, so remove the current runtime
hack we have.
Alexander Graf [Thu, 25 Jan 2018 11:05:42 +0000 (12:05 +0100)]
serial: Use next serial device if probing fails
Currently our serial device search chokes on the fact that the serial
probe function could fail. If it does, instead of searching for the next
usable serial device, it just quits.
This patch changes the fallback logic so that even when a serial device
was not probed correctly, we just try the next ones until we find one that
works.
Signed-off-by: Alexander Graf <agraf@suse.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Faiz Abbas [Wed, 24 Jan 2018 09:14:49 +0000 (14:44 +0530)]
am33xx: board: Call spl_early_init() to support sdram_init()
With driver model enabled in SPL, sdram_init() requires device tree
and malloc to be initialized.
Therefore call spl_early_init() in early_system_init().
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com> Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Alexander Graf [Tue, 23 Jan 2018 17:05:22 +0000 (18:05 +0100)]
mmc: Add bcm2835 sdhost controller
The BCM2835 family of SoCs has 2 different SD controllers: One based on
the SDHCI spec and a custom, home-grown one.
This patch implements a driver for the latter based on the Linux driver.
This is needed so that we can make use of device trees that assume driver
presence of both SD controllers.
Alexander Graf [Tue, 23 Jan 2018 17:05:21 +0000 (18:05 +0100)]
bcm283x: Add pinctrl driver
The bcm283x family of SoCs have a GPIO controller that also acts as
pinctrl controller.
This patch introduces a new pinctrl driver that can actually properly mux
devices into their device tree defined pin states and is now the primary
owner of the gpio device. The previous GPIO driver gets moved into a
subdevice of the pinctrl driver, bound to the same OF node.
That way whenever a device asks for pinctrl support, it gets it
automatically from the pinctrl driver and GPIO support is still available
in the normal command line phase.