When using MbedCRC, init value must be non-reversed, regardless of
`reflect_data` or `reflect_out` settings. This means we need to reflect
the intermediate output before passing using it as the next init value.
(In GCC this ends up putting in two `RBIT` instructions back-to-back,
because it's implemented as assembler, so it doesn't know how to
optimise. In ARMC6, `__RBIT` is implemented as an intrinsic, so adding
this reflection cancels the existing reflection and makes the code
smaller).
* Make mbed_error use bitwise MbedCRC call rather than local
implementation.
* Remove use of POLY_32BIT_REV_ANSI from LittleFS.
* Move some MbedCRC instances closer to use - construction cost is
trivial, and visibility aids compiler optimisation.
* Use compile-time detection of hardware CRC capability, so unneeded
code and tables do not go into the image.
* Add global JSON config option to allow choice between no tables,
16-entry tables or 256-entry tables for software CRC. Default set
to 16-entry, reducing ROM size from previous 256-entry.
* Allow manual override in template parameter to force software or
bitwise CRC for a particular instance.
* Micro-optimisations, particularly use of `RBIT` instruction and
optimising bitwise computation using inline assembler.
Incompatible changes:
* Remove special-case "POLY_32BIT_REV_ANSI" - users can use standard
POLY_32BIT_ANSI, which now uses the same 16-entry tables by default,
or can use hardware acceleration, which was disabled for
POLY_32BIT_REV_ANSI. MbedCRC<POLY_32BIT_ANSI, 32, CrcMode::TABLE> can
be used to force software like POLY_32BIT_REV_ANSI.
* The precomputed table for POLY_16BIT_IBM had errors - this has been
corrected, but software CRC results will be different from the previous
software calculation.
* < 8-bit CRC results are no longer are shifted up in the output value,
but placed in the lowest bits, like other sizes. This means that code
performing the SD command CRC will now need to use `(crc << 1) | 1`,
rather than `crc | 1`.
* Change "is supported" check to be a macro, so it can be done at
compile-time.
* Eliminate weird shift on 7-bit CRCs.
* Add support for 32-bit CRCs and reversals to TMPM3HQ.
As the `psa` library is not included in the baremetal profile, perform
a TFM system reset only if the `psa` library is included in
the build otherwise perform a normal CMSIS system reset.
If the fault handler was hit before the stdio console was used and
initialised, the initialisation code caused a "mutex in ISR" trap,
stopping the register dump from happening.
Temporarily set the `error_in_progress` flag at the top of the fault
handler, and restore it before calling `mbed_error`. Take the
opportunity to suppress fault dumps on recursive crashes, much as is
done inside `mbed_error`.
Previously Greentea tests was not initialising its storage
before asking for bd->get_program_size(), causing FlashBlockDevice to
return zero. This caused both TDBStorage's to use zero for both
parameter to SlicingBlockDevice(bd, 0, 0), effetivaly both then
used same addresses for slice. This caused SecureStore tests
to fail, because writes to internal RBP storage overwrote keys
from external storage.
Fine-tune TDBStore sizes, so that all tests can fit into storage.
At least with LPC55S69's default TDBStore configuration it's
impossible to run storage Greentea tests without exhausting the
memory reserved for storing keys.
Fixes an issue where number of keys were removed based on number of
threads which didn't have anything to do with the test case.
Fixes an issue where number of keys were assumed to be constant
but variable number was used for configuration.
In case our are contains data from previous reset() or reset_area(),
we might end up in the situation where free space contains valid
key headers, but we have not erased that area yet. This can cause
failures if the deinit() and init() because new scan of that area
would continue as long as keys are found. This causes keys on the
not-yet-erased area to be included in the new instance of TDBStore.
To prevent this failure, check after each key-write that our free
space does not contain valid key headers. Also make sure that we
erase one program unit sector over the master record. If we erased
just the master record,first key might is still there, causing next
init() to find it. Extend erase area by one program unit, so that
build_ram_table() won't find any keys.
Kevin Bracey
Veijo Pesonen
Martin Kojtal
Qinghao Shi
Ari Parkkila
Hamza Rizwan
jeromecoutant
Michal Paszta
Hugues Kamba
Evelyne Donnaes
Piotr Stolarz
Seppo Takalo
Przemyslaw Stekiel
Antti Kauppila