The original size given to the "calloc" call did not account for the
necessity to pad section sizes to a multiple of 4 bytes. As such, every
time we wrote a section that did not have a size multiple of 4 we would
overflow the allocated buffer.
Fix the calloc call to account for that padding requirement.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
The SRC should be a normal LL component, it can process 1 ms blocks
of audio in real-time. The domain_types = "1" is for DP scheduling
that is not enabled currently in Zephyr configuration. It causes a
pipeline prepare fail when a topology contains SRC.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
The deleted "install" target copied only the rimage executable and
left the config files behind. This gave the wrong impression that the
executable is useful without config files.
"Installing" also gave the wrong impression that rimage versions are
somewhat stable and relatively independent of SOF versions: they're
not. In fact there is no such thing as an rimage "version": everyone
should always use the exact rimage _git commit_ from the west manifest
or git submodule. There are no rimage "releases" and no semantic
versioning or anything like it, rimage is effectively part of the SOF
source and build and run directly from its build directory by
practically every SOF developer and SOF CI thanks to
sof/src/arch/xtensa/CMakeLists.txt#ExternalProject_Add(rimage_ep ...
sof/scripts/xtensa-build-zephyr.py#def build_rimage()
sof/zephyr/CMakeLists.txt#set(RIMAGE_CONFIG_PATH ...
etc.
Providing an "install" target greatly increases the chances of
different people and CIs using different rimage versions which is the
last thing we want considering the many significant rimage changes
happening right now, examples:
- https://github.com/zephyrproject-rtos/zephyr/pull/56099
- https://github.com/zephyrproject-rtos/zephyr/pull/54700
- https://github.com/thesofproject/sof/issues/7270
- https://github.com/thesofproject/sof/issues/7304
- https://github.com/thesofproject/sof/pull/7320
- https://github.com/thesofproject/rimage/pull/155
While it's useful for multiple files (e.g.: config files), a CMake
target was always overkill to copy a single file. Someone or some
script who really wants to copy the rimage binary to some other place
that the build directory for some (discouraged) reason _can still do
so_ with a much more basic, simpler and more transparent file copy
command.
Finally, the default "bin" DESTINATION required root access which
is otherwise totally unncessary to build SOF.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Some platforms (eg. tgl) dont have modules configuration in a toml files.
An attempt to use the configuration template ended with a reference to
a null pointer. This commit fixes this issue.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
The manifest for elf modules is read from their .module section.
The modules configuration parsed from a toml file was completely ignored.
After this change, the parsed configuration is treated as a template for
the module manifest.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Currently, sof incorrectly includes a private manifest.h header belonging
to the manifest.c module. Moved the MAN_MAX_SIZE_V1_8 definition used by
sof, to the public manifest header to allow sof to switch to using the
correct file.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
A new -l parameter has been added to the program's command line. It
instructs the program to not treat the first module as a bootloader
and not to skip the bss section for it.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
A is_bootloader field has been added to determine whether a module is
a bootloader. This is clearer than checking the module index and the number
of modules in different places in the program.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Changing the name of the structure type is a prelude to separating
information about the elf file from the information used during creation
of a manifest.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
A new memory_alias structure containing memory address alias configurations
has been created. It was placed in the memory_config structure.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Added name_val structure representing name/value tuple. Added function
print_enum which displays the name of the enum corresponding to a given
value. Added print_flags function which display names of flags constituting
a given value.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
There was no dram_offset value in any configuration file. It also wasn't
used anywhere in the code. So the code reading it from the tolm file has
been removed.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
The man_create_modules function may return an error that was ignored. Added
check of a return value.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
cAVS 1.5, 1.8 and 2.0 are supported by Zephyr, their configurations
need alias definitions.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Now that rimage can distinguish between Zephyr and XTOS builds, we
can use memory alias specifiers with both.
This reverts commit 1e0a85b44a.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
The fseek function returns -1 to signal an error. Return value was assigned
to a variable without a sign, so the error could not be detected.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Added error handling where it was missing. According to the documentation,
the fseek function returns a non-zero value on error. The conditions for
checking this value have been corrected.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Fields related to hash computing have been removed from the image
structure. Removed openssl headers from rimage.h, which made it necessary
to add some missing includes in several files.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Switched to use new hash functions. The parameter list for the rimage_sign
and rimage_verify functions has been simplified. The necessary information
is now included in the hash_context structure. In addition, code fragments
that may be common to different versions of the manifest have been
separated.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Prepared a new set of functions for computing digest with error handling.
The hash context has been placed into a separate structure. This allows
to conveniently pass the calculated digest along with its length and used
algorithm.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
The version fields in the fw_ext_man_cavs_header structure describe version
of the structure itself, not the image. The correct version number has been
set.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
Modified the strtoul function parameter to automatically recognize
the value format. This will avoid the situation when a decimal value
is parsed as hexadecimal.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
A new function create_file_name has been created that prepares the name
of the output file with the given extensions. Unlike the previous solution,
it checks whether the new name will fit in the output buffer.
A new file_utils.c file has been prepared, into which generic functions to
simplify the use of files will be moved.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
A set of functions which make it easier to parse toml files has been
separated into a new file.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
In image re-signing mode or if there are no input files, the memory
allocated for heap_adsp was not released.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
The init_config is used to indicate if base config extension is
used (1) or not (0). Its default value is used if not specified
in module config, the default value for init_config should be zero
instead of one.
Signed-off-by: Chao Song <chao.song@linux.intel.com>
We don't include ROM sections into the output image. If conversion is
applied to them, they end up in the SRAM address range, which then
generates corrupted images.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Paul Olaru