| Age | Commit message (Collapse) | Author |
|---|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
the time difference used here could go negative, which
would overflow in the xor op on mix, leading to a biased
entropy pool. we want to ensure that they numbers do
not overflow, because here they are cast to unsigned
which would then produce very large numbers.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
not at runtime
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
i already send the right arg anyway. this is a
preventative bug fix against future maintenance.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
already checked below, then err()
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
zero never occurs, because rw_file_once never returns zero,
but only rw_file_once determines that. rw_file_exact must
handle every possible error.
right now, if that call returns zero, rw_file_exact would
have an infinite loop.
this doesn't actually happen at the moment, so this is a
preventative bug fix.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
for fussy static analysers and/or compilers
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
properly verify the value of the arguments, with
asserts.
add simpler runtime checks in-function, on prw,
rw_file_once and rw_file_exact.
variable names in english now, and the code is
cleaner, while being functionally equivalent.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
useless check. will never be true.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
itt's totally obvious
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
for pulses, we currently use amplitude detection.
edge detection is better, because weak / low gain
signals will be more reliable. if audio is coming
in on/from a system that does automatic gain
adjustment, this once again is more robust too.
microphones and speakers (which people often use
with spkmodem if nothing else available) often
clamp amplitude, to an extent that this software
may not detect those pulses reliably that way.
so we detect slope edges instead. this causes
very little performance penalty (use of abs(),
that's about it)
however, edge detection is inherently vulnerable
to noise, so we will also detect amplitude. this
acts as an effective noise filter, while still
improving pulse detection.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
in select_separator_tone, i never reset these
after computing their average.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
this check no longer applies (never triggers)
is_signal_valid already guarantees that the separator
tone is valid.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
otherwise, calibration could collect garbage data.
this improves noise mitigation.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
bits are currently assembled even on invalid frames. this
patch fixes that - the bug is also in the GNU version.
this reduces the chance of noise/calibration from creating
corrupt character output during operation.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
improves reliability in the case when audio cuts out,
mic glitches, laptop audio power saving, etc.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
enforce at least two tones. this mitigates the
chance of random noise being treated as a real
tone, and reduces the chance of broken
thresholds versus freq min/max e.g.
freq min 31, max 32 and threshold 31
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
we weren't actually using what we calculated. this patch
fixes that, thus preventing random noise / microphone
clicks, random artifacts and such from being treated
as real frames (the purpose of is_valid_signal is
partly noise suppression).
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
in handle audio, i do the number of samples
per frame, and one more. e.g. 241 instead of
240. this bug is in the original GNU version
too. this patch fixes it.
this means that the output could slowly go
out of sync with calculated timings. the
patch fixes that. in practise, the decoder
is not that sensitive, and the code would
adjust anyway (automatic timing adjustment),
but ideally we want to not *cause* such
issues even if we mitigate them.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
This reverts commit dbf0c3ccc276eeb7e4d80efe75bdd684cc3c6444.
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
assert integer sizes, important in this program because
we make several implicit assumptions about word sizes,
and integers need to be of a certain size.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
collect_separator should be select_separator, to bring
it in line with select_low_tone
this just makes the code a bit easier to read
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
i accidentally left this reset here during a
previous refactor.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
auto seems redundant. detect implies auto.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
the fir filter produces stable frequencies per frame,
but learning per sample (within a frame) means we
record the same value roughly 240 times.
here, we are syncing up at just the right moment
instead, and only at that moment, this increasing
both performance and reliability.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
oops!!!
another mistake during refactoring. right now it
doesn't increment before being checked, so learning
can go forever in an infinite loop
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
oops
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
i conflated two separate tests in a previous change.
the silence check was defeated by still checking f
alongside it, which would be set, thus satisfying
the condition, and proceeding, which defeats the
purpose of the silence check (ignore false signal
that is actually noise) - so in the original patch
that i wrote, the extra checks actually do nothing.
this patch fixes that, and makes the logic a bit
clearer.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
otherwise, it runs twice
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
this enables the separated code to have reduced
indentation
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
the tone detection currently only tracks data, not
the separator. track both instead, for improved
detection reliability.
e.g.
separator tone ≈ 9
data low tone ≈ 18
data high tone ≈ 24
two fir windows produce e.g.
freq data 9 sep 0
then 18, 9
then 24, 9
18, 9
so we take min(data, separator)
that gives 9,9,9,9
now we have the separator cluster
however, if both windows are active during transitions,
you can also capture the higher clusters, which would
allow freq_max to grow
so when you learn e.g.:
freq min = 9
freq max 24
then the learned threshold would be:
(9 + 24) / 2 = 16
and now you know how to separate the tones
fir already suppresses noise so the pulse should
be reliable. so freq/sep only go non-zero when an
actual tone exists
this should now result in being able to sync with
spkmodem encoders with no prior knowledge of the
correct tone frequences. we just use maths.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
since we have auto-detection now, we only need to know
that two signals exist, not that they are valid, since
the auto-detection now handles validation and fallback.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
a continuation of the previous patch. this waits for
currently one second, before defaulting to the hardcoded
value. otherwise, it tries to use whatever timing it
gets automatically.
this way, the program should now reconfigure its own
timing, without intervention by the user, if the timing
differs from sensible defaults.
this is because spkmodem is implementation-defined;
it's just however coreboot and/or GRUB happen to set
it up, and on the hardware in question.
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
current logic is hardcoded, as in the original spkmodem-recv.
with this change, small differences are observed and averaged,
then the detection thresholds are adjusted accordingly.
the existing macros serve as a baseline, but real signals
differ. with this change, we therefore account for possible
drift in timings, which can change in real-time; the old
code could possibly get out of sync beccause of that, which
may have resulted in corrupt characters on the screen. this
change therefore should make the output a bit more stable.
the detection window is continually adjusted, so that the
output timings don't drift.
the tolerances are automatically adjusted based on base
timings (see new define in patch)
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
|
make the frequencies clearer in printf
Signed-off-by: Leah Rowe <leah@libreboot.org>
|
