dsp: base: add encode_pldm_header_only_errno()

Currently, `encode_pldm_header_only()` returns PLDM Completion Code,
which is deprecated, but most of this API's use case were internally
inside libpld

dsp: base: add encode_pldm_header_only_errno()

Currently, `encode_pldm_header_only()` returns PLDM Completion Code,
which is deprecated, but most of this API's use case were internally
inside libpldm itself, therefore, add `encode_pldm_header_only_errno()`
as an internal API.

Change-Id: I87822a4f6afe8aa8eb87034179c37341d7ca4190
Signed-off-by: Unive Tien <unive.tien.wiwynn@gmail.com>

show more ...

Apply GCC's tainted_args attribute to library entrypoints

The implementation applies `__attribute__((tainted_args))` by
integrating it into the existing ABI macro annotations.

In the process, quite

Apply GCC's tainted_args attribute to library entrypoints

The implementation applies `__attribute__((tainted_args))` by
integrating it into the existing ABI macro annotations.

In the process, quite a number of APIs were discovered to be unsafe in
ways that were not immediately fixable. Often this is because they lack
arguments that enable the appropriate bounds-checking to be applied.

Redesigning them is work beyond the scope of the immediate
effort. Instead, we also introduce a new annotation,
LIBPLDM_ABI_DEPRECATED_UNSAFE, that simply lacks
`__attribute__((tainted_args))` and therefore doesn't trigger the extra
analysis.

Change-Id: Ib8994eaa3907a5432d040426ad03687cbf4c2136
Signed-off-by: Andrew Jeffery <andrew@codeconstruct.com.au>

show more ...

dsp: base: Rework {un,}pack_pldm_header() error handling

The current preference is that library APIs return negative errno values
to signal implementation errors. That doesn't jive with existing sta

dsp: base: Rework {un,}pack_pldm_header() error handling

The current preference is that library APIs return negative errno values
to signal implementation errors. That doesn't jive with existing stable
APIs returning PLDM completion codes, so provide a means to translate
between the two.

The first users are the {un,}pack_pldm_header() functions.

Change-Id: I7b7cb97a1d8b96ec0fec1c0a5fbd8503da834d86
Signed-off-by: Andrew Jeffery <andrew@codeconstruct.com.au>

show more ...

libpldm: Reorganize source and test files

Primarily this is about moving specification-specific files into
'dsp/' (in the "DMTF Standard Publication" sense[1]) subdirectories of
both src/ and tests/

libpldm: Reorganize source and test files

Primarily this is about moving specification-specific files into
'dsp/' (in the "DMTF Standard Publication" sense[1]) subdirectories of
both src/ and tests/.

[1]: https://www.dmtf.org/sites/default/files/standards/documents/DSP4014_2.14.0.pdf

libpldm is a concrete C implementation of the PLDM family of
specifications. This invokes some accidental complexity[2] such as the
msgbuf APIs and other concerns.

[2]: https://en.wikipedia.org/wiki/No_Silver_Bullet

Separate the essential complexity (everything under the dsp/
subdirectories) from the accidental complexity (almost everything else).

While doing so, I took the opportunity to drop the 'libpldm_' prefix
and '_test' suffix from a variety of tests. The 'libpldm_' prefix is a
hangover from the days when libpldm was a subproject of OpenBMC's pldm
repo. The '_test' suffix feels redundant given the parent directory
path.

Note that we maintain separation of the src/ and tests/. The test suite
is implemented in C++ while libpldm's APIs are declared and defined in
C. The ability to chop all the tests and C++ out of the implementation
by ignoring a subtree seems like a desirable property when vendoring the
library into other projects.

Finally, update the x86_64 GCC ABI dump, as rearranging the source
causes a lot of churn in its definitions.

Change-Id: Icffcc6cf48b3101ecd38168827c0a81cffb8f083
Signed-off-by: Andrew Jeffery <andrew@codeconstruct.com.au>

show more ...