This was a small hole in the logic that handles initializing the
lookahead buffer. To imitate exhaustion (so the block allocator
will trigger a scan), the lookahead buffer is rewound a full
lookahead and set up to look like it is exhausted. However,
unlike normal allocation, this rewind was not kept aligned to
a multiple of the scan size, which is limited by both the
lookahead buffer and the total storage size.
This bug went unnoticed for so long because it only causes
problems when the block device is both:
1. Not aligned to the lookahead buffer (not a power of 2)
2. Smaller than the lookahead buffer
While this seems like a strange corner case for a block device,
this turned out to be very common for internal flash, especially
when a handleful of blocks are reserved for code.
- Add flag to SPI class to track if the SPI instance has locked deep sleep mode.
- Wrap call to sleep_manager_lock_deep_sleep to only be called if SPI instance
hasn't already locked deep sleep.
- Wrap call to sleep_manager_unlock_deep_sleep to only be called if SPI has
currently locked deep sleep mode.
- Add flag to I2C class to track if the I2C instance has locked deep sleep mode.
- Wrap call to sleep_manager_lock_deep_sleep to only be called if I2C instance
hasn't already locked deep sleep.
- Wrap call to sleep_manager_unlock_deep_sleep to only be called if I2C has
currently locked deep sleep mode.
This bug was a result of an annoying corner case around intermingling
signed and unsigned offsets. The boundary check that prevents seeking
a file to a position before the file was preventing valid seeks with
positive offsets.
This corner case is a bit more complicated than it looks because the
offset is signed, while the size of the file is unsigned. Simply
casting both to signed or unsigned offsets won't handle large files.
which can be called from diferent contexts.
Orginal nordic critical primitives must been called in pairs
from exacly the same function. As mbed hal call it in separate
methods, so they are not suitable here.
Parsing dispatch now uses a metaclass as a Trait and a series of classes
that implement the trait for dispatching parsing. This structure gives
each parser it's own namespace, avoiding tacking on suffixes to each
attribute indicating which parser uses it.
As it does not share the codebase with RZ A1H, it needs to be disabled until
bring it up to date with the latest changes to cmsis. There are changes
regarding caches, mmu and others
Build api used to build cmsis separately (how mbed 2 is being build). This is
currently not how cmsis is being defined. As there target dependencies in some
cases, we should include paths from targets when building cmsis
The API is as per posix standard, but does not provide stats for file/directory.
Stats buffer (block size, total block count, free block count) is filled for
entire mounted filesystem.
Merged from the existing TARGET_DISCO_F413ZH code (which has the same MCU)
and TARGET_NUCLEO_F412ZG code (which has the same Nucleo-144 board and
pinout).
Jimmy Brisson
Christopher Haster
bcostm
Martin Kojtal
Steven Cartmell
jeromecoutant
Andrzej Puzdrowski
Anna Bridge
gorazd
Deepika
Adam Heinrich