common: ubsan: ignore shifting one into sign bit ...

The __ubsan_handle_shift_out_of_bounds handler would be called for code
shifting a one into the sign bit like (1 << 31), which is all too common
in barebox.  It's technically UB, but it's so prevalent that it's highly
unlikely to be treated by a compiler as anything else than the
standard-compliant (1U << 31).

Check for this case here and ignore it selectively.

Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

trivial: fix typo lenght -> length ...

Generated by

	perl -p -i -e 's/lenghte?/length/' arch/arm/boards/chumby_falconwing/falconwing.c arch/arm/mach-samsung/mem-s3c64xx.c fs/ext4/ext_common.h lib/gui/lodepng.c include/jtag.h

There is another instance in dts/Bindings/usb/s3c2410-usb.txt, this is
fixed in Linux v5.3-rc1 so supposed to be fixed soon in barebox, too.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Convert latin1 files to utf-8 ...

The source files are parsed by $(make docs). Currently this is done
using python2 (actually 'python' which is still python2 on most
machines). When using python3 however (which has a stricter encoding
handling) it is easier if all input is in utf-8 encoding. So convert the
few remaining latin1 encoded files to utf-8. This only affects code
comments.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

libfile: Document that read_file() returns a terminated buffer ...

read_file() and read_file_2() return a '\0' terminated buffer to make
sure the buffer is usable as a string. Other code like read_file_line()
depends on this behaviour, so document it explicitly.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/qemu'

Merge branch 'for-next/misc'

common: add generic CONFIG_UBSAN plumbing ...

-fsanitize=undefined allows compile-time instrumentation of code to detect
some classes of runtime undefined behavior.

In preparation for allowing arches to provide infrastructure in support
of this feature, add some generic UBSAN options and associated plumbing.

These are only shown in the debug menu when the arch selects the
appropriate symbol. The option is named equally to their Linux counterparts.

Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libfile: fix typos ...

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

lib: add HAVE_EFFICIENT_UNALIGNED_ACCESS Kconfig option ...

The LZO, LZ4, DEFLATE implementations in lib/ already check against
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS. Define this symbol defaulting to
0. This makes porting Kconfig symbols depending on it easier and in future,
we could select it from different arches.

Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Kconfig: create Kconfig symbol for ARCH_HAS_DATA_ABORT_MASK ...

Other arch-specific features are exposed in Kconfig too, so do here
likewise.

Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Kconfig: create Kconfig symbol for ARCH_HAS_STACK_DUMP ...

Other arch-specific features are exposed in Kconfig too, so do here
likewise.

Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libfile: have write_full return a descriptive error code ...

So far (p|)write_full has been returning -1 on error.
Some callers of write_full like imx_bbu_write_device print out the
function's return value on error, effectively rendering every printed
error message to be due to EPERM.

On the other hand, some callers like do_memcpy, use it correctly and
use errno. Sidestep the issue by having the function return -errno on
errors as well.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

gui: png_lode: fix freeing of uninitialized pointer ...

If either calloc or png_uncompress_init fails, free(png) will free the
uninitialized png pointer. Avoid this and while at it postpone the
img allocation till after the early exit.

Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Compile libfdt for barebox ...

libfdt can be useful for very early device tree parsing, so compile it
for barobox aswell. This is done the same way as done in the Kernel:
we put C files with the same name as in libfdt into lib/ and just
include the libfdt files from there.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

lib: Add missing arguments to memcpy_parse_options() ...

Memcpy use-case differs from that of memcmp in default access type and
destination file mode. This was missed in original commit that
introduced memcpy_parse_options(). Add said parameters to
memcpy_parse_options(), so the can be correctly specified depending on
the use-case.

Fixes: ddf4cca339 ("commands: Introduce memcpy_parse_options()")
Reported-by: Peter Mamonov <pmamonov@gmail.com>
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/misc'

commands: Introduce memcpy_parse_options() ...

Both memcpy and memcmp have identical options, so in order to share
code between them, introduce memcpy_parse_options() and change both
tools to use it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libbb: Drop simple_itoa() ...

Hush is the only one user of simple_itoa() and the code there can be
re-implemented using snprintf(). Change the code to get rid of
simple_itoa().

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libfile: Do not return 0 from write_full() ...

None of the callers of write_full() expect a zero return value. Given
how the documentation explicitly states that either all of the buffer
is going to be written out or an error generated, treat 0 retrun from
write() as a error, set errno to ENOSPC and return -1.

Same logic applies to pwrite_full() as well, so make the change there
while at it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libfile: Simplify read_full() ...

We can figure out the amount of written data by substracting 'insize'
from 'size' so there is no need to keep a separate counter for that.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

lib: select BASE64 if BLOBGEN is selected ...

Blobgen requires the BASE64 libraries, select them to provide them if
BLOBGEN is selected.

Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

lib: add blobgen framework ...

This adds a framework for en/decrypting data blobs. Some SoCs have
support for hardware crypto engines that can en/decrypt using keys
that a tied to the SoC and are visible for the crypto hardware only.
With this patch it's possible to encrypt confidential data using
these keys and to decrypt it later for usage.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

lib: add base64 helpers ...

Import the busybox-1.23.1 uuencode/base64 helper functions.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/misc'

Merge branch 'for-next/lseek'

treewide: surround Kconfig file paths with double quotes ...

Based on Linux commit 8636a1f

This will be needed when you sync Kconfig with Linux 5.0 or later.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Add Freescale QUICC Engine firmware support ...

The Freescale QUICC Engine found on Layerscape SoCs needs firmware
loaded. This adds support for loading such a firmware.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libfile: Make use of write_full() in copy_file() ...

Change copy_file() to rely on write_full() instead of re-implementing
it locally.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

libfile: Make failure path of open_and_lseek() consistent ...

Change the code of open_and_lseek() to make sure that opened file is
closed in every failure path as well. While at it make sure we don't
mix returning function return code and returning errno by opting to
always return the former.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/spd'