normally we never see these failures, but, as we are seeing
flash wear-out on the CI machine the writes can take
long enough that we get some concurrent requests. As a work-around,
just automatically retry a few times before throwing the error
back to the caller.
The rationale is that xous-rs should have only architecture
specific code, e.g.: riscv, arm, x86, etc. To make other
crates that depend on it not specify the features.
This could be useful when using xous-rs with `libstd`
for example.
Signed-off-by: Jean-Pierre De Jesus DIAZ <me@jeandudey.tech>
In general, hosted mode is assumed "by default" if the target
is not Xous. This allows packages to be built cleanly on command
lines that don't know to pass a --feature flag to specify a target.
However, an explicit "hosted" flag is still retained so that we
can confirm the build system intent. This saves us from the failure
mode that we thought we were building for target A, but accidentally
ommitted the flag, and instead we've been silently building for
"hosted" mode without any warning. The main consumer of the "hosted"
flag is now just the build.rs script inside of the utralib.
suspect maybe the custom vex routine for cache flush has a subtle
issue, so fall back to first-principles read-region flush.
remove excess calls to flushing as well, turns out we effectively
duplicated the call because it was both in the core patch routine
and outside of it too.
after updating a piece of FLASH memory, there is an inherent
cache coherence issue. This is because the write happens through
a side channel that is outside of the memory model of the system:
writes happen through a CSR that feed data to pages of FLASH.
Reads happen through direct-mapped memory regions that are unaware
of write mutations.
Thus, after writing a page of memory, if you try to access it,
sometimes you'll read the old cached value instead of the new
updated value.
This patch introduces a mandatory dcache flush after every
FLASH write, so that any cached FLASH regions are re-fetched.
This will allow for e.g. xous on riscv32-imac but not a precursor
target, for example: by using feature selectors we get to have
a bigger range of targets that we support, such as for implementations
that haven't been upstreamed or have official target-triples.
The downsides are:
- patch override libraries don't absorb feature flags. so, for
example, the AES crate that is used to patch other crates won't
be aware of the target features without modifying the referencing
crates.
- features are normally meant to be additive, but target features
are mutually exclusive. The mutex is enforced by a check in
utralib that throws a panic on build if two targets are accidentally
selected at the same time.
