18 May, 2017
2 commits
-
This patch updates the mksunxiboot tool to optionally add
the default device tree name string to the SPL header. This
information can be used by the firmware upgrade tools to
protect users from harming themselves by trying to upgrade
to an incompatible bootloader.The primary use case here is a non-removable bootable media
(such as NAND, eMMC or SPI flash), which already may have
a properly working, but a little bit outdated bootloader
installed. For example, the user may download or build a
new U-Boot image for "Cubieboard", and then attemept to
install it on a "Cubieboard2" hardware by mistake as a
replacement for the already existing bootloader. If this
happens, the flash programming tool can identify this
problem and warn the user.The size of the SPL header is also increased from 64 bytes
to 96 bytes to provide enough space for the device tree name
string.
[Andre: split patch to remove OF_LIST hash feature]Signed-off-by: Siarhei Siamashka
Signed-off-by: Andre Przywara -
mksunxiboot limits the size of the resulting SPL binaries to pretty
conservative values to cover all SoCs and all boot media (NAND).
It turns out that we have limit checks in place in the build process,
so mksunxiboot can be relaxed and allow packaging binaries up to the
actual 32KB the mask boot ROM actually imposes.
This allows to have a bigger SPL, which is crucial for AArch64 builds.Signed-off-by: Andre Przywara
Reviewed-by: Simon Glass
Reviewed-by: Jagan Teki
12 May, 2017
2 commits
-
As per MAINTAINERS[1] file description, 'Supported' is
a valid status for a board. But buildman thinks 'Maintained'
is the only valid state and complains about boards with 'Supported'
status. Update buildman to accept 'Supported' as valid state.[1] http://git.denx.de/?p=u-boot.git;a=blob;f=MAINTAINERS;h=0962b47bf9057b22e93624e070c0204b893790dc;hb=HEAD#l10
Reported-by: Sekhar Nori
Signed-off-by: Lokesh Vutla
Reviewed-by: Tom Rini
Reviewed-by: Simon Glass -
The rsa_st struct has been made opaque in 1.1.x, add forward compatible
code to access the n, e, d members of rsa_struct.EVP_MD_CTX_cleanup has been removed in 1.1.x and EVP_MD_CTX_reset should be
called to reinitialise an already created structure.Signed-off-by: Jelle van der Waa
11 May, 2017
10 commits
-
At present dtoc assumes that nodes which are phandles do not themselves
reference other phandle nodes. Unfortunately this is not necessarilly
true. As a result we can currently output C code which does not compile
because a node declaration can be referenced before it is declared.Adjust the code to explicitly output all phandle nodes needed by node
before the node itself is output.This fixes building with the latest rk3399-firefly.dts from Linux, which
has reordered the nodes.Signed-off-by: Simon Glass
Tested-by: Kever Yang -
The code to generate the tables is quite long. Move the node-output code
into its own function.Signed-off-by: Simon Glass
Tested-by: Kever Yang -
The imagetool framework checks whether function pointer for the verify,
print and extract actions are available and will will handle their
absence appropriately.This change removes the unnecessary functions and uses the driver
structure to convey available functionality to imagetool. This is in
fact better than having verify just return 0 (which previously broke
dumpimage, as dumpimage assumed that we had handled the image and did
not continue to probe further).Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
Dumpimage (it invoked with "-T rkspi" or "-T rksd") would not work due
to check_params failing. These changes ensure that we can both be called
with an empty imagename.Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
This change set adds documentation to the header0 initialisation and
improves readability for the calculations of various offsets/lengths.As the U-Boot SPL stage doesn't use any payload beyond what is covered
by init_size, we no longer add RK_MAX_BOOT_SIZE to init_boot_size.Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
The calculation of the variable header size in rkcommon_vrec_header
had been update twice in the earlier series (introducing boot0-style
images to deal with the alignment of the first instruction in 64bit
binaries). Unfortunately, I didn't update the comment twice (so it
remained out-of-date).This change brings the comment back in-sync with what the code is
doing.Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
In (first) breaking and (then) fixing the rkspi tool, I realised that
the calculation of the required padding (for the header-size and the
2K-in-every-4K SPI layout) was not as self-explainatory as it could
have been. This change rewrites the code (using new, common functions
in rkcommon.c) and adds verbose in-line comments to ensure that we
won't fall into the same pit in the future...Tested on the RK3399 (with has a boot0-style payload) with SD/MMC and SPI.
Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
Our earlier change broke the generation of SPI images, by excluding the
2K used for header0 from the size-calculation.This commit makes sure that these are included before calculating the
required total size (including the padding from the 2K-from-every-4K
conversion).Signed-off-by: Philipp Tomsich
Acked-by: Simon Glass -
Add support for rk3328 package header in mkimage tool.
Signed-off-by: Kever Yang
Reviewed-by: Simon Glass
08 May, 2017
3 commits
-
The Kconfig description replaces the description in the README file so
as options are migrated they can be removed from the README.Signed-off-by: Chris Packham
Reviewed-by: Simon Glass -
After moving to KConfig and removing from all headers options should be
removed from config_whitelist.txt so the build starts complaining if
someone adds them back.Acked-by: Masahiro Yamada
Signed-off-by: Chris Packham
Reviewed-by: Simon Glass -
Avoid repetitive code dealing with asking the user for confirmation.
Signed-off-by: Chris Packham
Reviewed-by: Simon Glass
06 May, 2017
2 commits
-
The evaluation of option -c is incorrect:
According to the C99 standard endptr in the first strtol is always
set as &endptr is not NULL.
So the first part of the or condition is always true.
If all digits in optarg are valid endptr will point to the closing \0
and the second strtol will read beyond the end of the string optarg
points to.Signed-off-by: Heinrich Schuchardt
Acked-by: Boris Brezillon -
va_start must always be matched by va_end.
The problem was indicated by cppcheck.
Signed-off-by: Heinrich Schuchardt
28 Apr, 2017
1 commit
-
An assignment (of a value to itself) was left over (after removing and
addition from the line) from moving the common padding code into
rkcommon_vrec_header.This change removes this to avoid a spurious warning in static code
analysis (i.e. Coverity).Signed-off-by: Philipp Tomsich
Reported-by: Coverity (CID: 161418)
Reviewed-by: Tom Rini
18 Apr, 2017
3 commits
-
If realloc fails we should release the old buffer.
Signed-off-by: Heinrich Schuchardt
Reviewed-by: Tom Rini
Reviewed-by: Simon Glass -
Many toolchains for ARM use the 'gnueabihf' suffix rather than just
'gnueabi', so allow these to be used, but with a lower priority than
'gnueabi' ones.Cc: Simon Glass
Signed-off-by: Tom Rini
Reviewed-by: Simon Glass -
Recently, U-Boot removed support for these architectures.
Signed-off-by: Masahiro Yamada
15 Apr, 2017
1 commit
-
When writing out some of our results we may now have UTF-8 characters
in there as well. Translate these to latin-1 and ignore any errors (as
this is for diagnostic and given the githash anything else can be
reconstructed by the user.Reviewed-by: Simon Glass
Signed-off-by: Tom Rini
14 Apr, 2017
6 commits
-
Here with some DM changes as well as the long-standing AT91 DM/DT
conversion patches which I have picked up via dm. -
One of these has crept in in this commit:
40a808f1 ARCv2: SLC: Make sure busy bit is set properly on SLC flushing
Adjust buildman to handle it.
Signed-off-by: Simon Glass
-
Not force to use python from PATH. Issue was noted when building with
Yocto, because python from the distro is always taken instead of
python-native built during Yocto process.Signed-off-by: Stefano Babic
CC: Simon Glass
Reviewed-by: Simon Glass -
With skeleton.dtsi being dropped it is more likely that the /aliases node
will be last in the device tree. Update fdtgrep to handle this.Signed-off-by: Simon Glass
Tested-by: Masahiro Yamada -
With Python 3.5.2 encode will throw an exception if val is a byte array.
Decode it to a string first. This assumes it's utf-8, if it's not valid
utf-8 it will throw an exception.Signed-off-by: George McCollister
Acked-by: Simon Glass -
os.read() returns a byte array in Python 3.5.2 and needs to be converted
into a string. Check if the returned value is an instance of bytes and
if it is decode it as a utf-8 string. If it is not a utf-8 encoded string
the decoding may fail with an exception.Prior to this fix the comparisions check data == "" would fail when data
was b'' and would cause an infinite memory leaking loop. joins would
also fail with an exception below but due to the infinite loop it never
made it that far.Signed-off-by: George McCollister
Acked-by: Simon Glass
13 Apr, 2017
4 commits
-
fw_env_open allocates buffers to store the environment, but these
buffers are never freed. This becomes quite nasty using the fw_ tools as
library, because each access to the environment (even just reading a
variable) generates a memory leak equal to the size of the environment.Fix this renaming fw_env_close() as fw_env_flush(), because the function
really flushes the environment from RAM to storage, and add a
fw_env_close function to free the allocated resources.Signed-off-by: Stefano Babic
-
Changes in the environment library are difficult to tracked by programs
using the library. Add simply an API version number that must be
increased each time when the API is changed.This can be detected and a program can work with different versions of
the library.Signed-off-by: Stefano Babic
-
Move U-Boot private data into a separate file. This
lets export fw_env.h to be used by external programs
that want to change the environment using the library
built in tools/env.Signed-off-by: Stefano Babic
-
aes.h is a too generic name if this file can
be exported and used by a program.
Rename it to avoid any conflicts with
other files (for example, from openSSL).Signed-off-by: Stefano Babic
10 Apr, 2017
2 commits
-
With recent changes, some x86-specific rom tests of binman fail to
run. Fix it by adding missing filenames in corresponding entries.Signed-off-by: Bin Meng
Reviewed-by: Simon Glass -
Now that we have added file names from Kconfig in x86 u-boot.dtsi,
update binman to avoid using hard-coded names.Signed-off-by: Stefan Roese
Cc: Bin Meng
Cc: Simon Glass
Reviewed-by: Bin Meng
07 Apr, 2017
1 commit
-
This program generates raw SPL images that can be flashed on the NAND with
the ECC and randomizer properly set up.This has been copied (and tweaked to find the right headers) from the
sunxi-tools (https://github.com/linux-sunxi/sunxi-tools) upstream
repository, commit 1c3a6ca5.Signed-off-by: Maxime Ripard
Acked-by: Boris Brezillon
Reviewed-by: Hans de Goede
Reviewed-by: Jagan Teki
05 Apr, 2017
3 commits
-
To simplify the creation of AArch64 SPL images for the RK3399, we
use the ENABLE_ARM_SOC_BOOT0_HOOK option and prepend 4 bytes of
padding at the start of the text section. This makes it easy for
mkimage to rewrite this word with the 'RK33' boot magic.This change brings logic to calculate the header size and allocate
the header back in sync. For the RK3399 we now limit the header to
before the payload (i.e. the 'header0' and the padding up to the
actual image) and overwrite the first word (inserted by the
boot0-hook for this purpose) with the 'RK33' magic in-place.X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich
Tested-by: Klaus Goger -
The RK3399 boot code (running as AArch64) poses a bit of a challenge
for SPL image generation:
* The BootROM will start execution right after the 4-byte header (at
the odd instruction word loaded into SRAM at 0xff8c2004, with the
'RK33' boot magic residing at 0xff8c2000).
* The default padding (during ELF generation) for AArch64 is 0x0,
which is an illegal instruction and the .text section needs to be
naturally aligned (someone might locate a 64bit constant relative
to the section start and unaligned loads trigger a fault for all
privileged modes of an ARMv8)... so we can't simply define the
CONFIG_SPL_TEXT_BASE option to the odd address (0xff8c2004).
* Finally, we don't want to change the values used for padding of
the SPL .text section for all ARMv8 targets to the instruction
word encoding 'nop', as this would affect all padding in this
section and might hide errors that would otherwise quickly trigger
an illegal insn exception.To deal with this situation, we modify the rkimage generation to
- understand the fact that the RK3399 needs to pad the header to an
8 byte boundary using an AArch64 'nop'
- the necessary logic to adjust the header_size (which controls the
location where the payload is copied into the image) and to insert
this padding (AArch64 insn words are always little-endian) into
the image following the 4-byte header magic.X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich
Tested-by: Klaus Goger -
The RC4 encoding works on full blocks, but the calculation of the
starting offset and size are needlessly complicated by using a
reference value known to be offset into a block by the size of the
header and then correcting for the (hard-coded) size of the header
(i.e. 4 bytes).We change this over to use the RK_SPL_HDR_START directly (which is
known to be on a block boundary).X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich
Tested-by: Klaus Goger