| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Stack allocating the correlation output generates a call to the copy
constructor of an zero valued vector. We can avoid this extra copy
constructor with a pointer reference and dynamic allocation.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Allow non-in-place use of the delay setting. Internally, the delay call
creates a new vector and copies the contents back into the original.
Instead, provide the option to return the computed output vector
directly and remove an an extra copy in the process.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
On Beagle Board the call into the sinc() function is generating a lot of
load on the peak interpolation. Simplify the sinc() function with a
dedicated table lookup. Eventually, this table may be removed in favour
of using a precomputed filterbank for fractional delay determination.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Set a transceiver high level length value that specifies the largest
number of complex or real filter taps that we will encounter. This
allows preallocation of head room and prevents an extra allocation and
copy on every incoming receive burst.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
With testing on current UHD releases, currently 003.005.xxx series,
timeout errors on both receive and transmit are recoverable on network
and USB based devices. Remove the fatal error conditions.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Clock indications passed up to GSM core originate on the transciever
downlink side. Set priority to keep the flow of clock updates
consistent.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This includes ARM Cortex A8 and A15 powered device such as Beagle
Board, Gumstix driven E100 USRP, and Arndale board. Set the reduced
SPS value automatically for the user.
For x86, if we don't support SSE3, then the architecture is
probably ancient and not with using. Drop the sampling down anyways
to at least make an attempt. Non floating point SIMD devices (e.g.
Raspberry Pi) also fall in this category
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Preallocate and compute a bank of fractional sample delay filters.
The number of filters to allocate is specified by the DELAYFILTS
preprocessor definition with a default value of 64. The filters
themselves are sinc pulse generated with 20 taps and Blackman-harris
windowed .
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Break out the signalVector object and clean up the interface in the
process.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Currently the code allocations a signalVector and then copies
into a radioVector. This is unnecessary because the latter is
a derived class making the first allocation unnecessary.
Modify the radioVector constructor to allow direct use in the
case above.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This prevents the use of a copy constructor in the downlink
modulator and prevents a secondary memory allocation during
the convolution. Avoid both cases by dynamically allocating
with preloaded head room. The latter provides enough memory
before the first sample in the burst to cover the length
of the filter taps.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Complex-complex block multiples are used for phase rotation of
bursts. Optimization targeted from perf profiling.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
B2xx is a USB based device so use the USRP1 based adaptive flow
control window for transmit bursts. This adds additional stability
primarily on ARM platforms.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
The transceiver and underlying device drivers are threaded. use
the following priority levels.
0.50 - UHD driver internal threads
0.45 - Receive device drive thread
0.44 - Transmit device drive thread
0.43 - UHD asynchronous update thread (error reporting)
0.42 - Receive burst processing thread(s)
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Similar to the existing Intel SSE cases, add support for NEON vector
floating point SIMD processing. In this case, use ARM assembly
directly as the NEON intrinsics do not generate preferential code
output.
Currently support NEON vectorized convolution and floating point
integer conversions.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This patch primarily addresses devices with multiple RF front end
support. Currently device support is limited to UmTRX.
Vectorize transceiver variables to allow multiple asynchronous
threads on the upper layer with single downlink and uplink threads
driving the UHD I/O interface synchronously.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
There is no need to create this method. Just call the UHD interface
directly.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
USRPping and sigProcLibTest are in an unmaintained state,
while the intended functionality remains unknown. Stored
filter taps are also unused and should also be removed.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Collect the slot information into an indpendent state object. This
will allow us to easily create multiple instances of internal state
variables without having to replicate the transceiver object itself.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
For multiple transceiver connections, it is inappropriate to
allocate all sockets in the transceiver constructor due to not
knowing how many connections are avaialble in advance and for
error checking purposes. Instead, store the base socket address
port combination and setup the sockets in the initialization
call.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
The current status and operability of this compile option is
unknown. Remove due to lack of use, demand, and maintenance.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
We do not use subversion. We do not need to refer to subversion
reference numbers.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Underruns are only explicitly set on the downlink side. Overruns
are logged but unused. In either case, reset indicators to false
to avoid sending false state information.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Using non-SSE4.1 enabled architecture would cause undefined
reference to 'convert_si16_ps' call.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Currently the default configuration is to not build the full
transceiver, which is pointless. Set the UHD driver, which
includes either Ettus or Fairwaves variants, as the default.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Move x86 specific files into their own directory as this
area is about to get crowded with the addition of ARM
support.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Use the same measurement method for RSSI as the noise level. Previous
method was to use the peak correlation amplitude relative to the
expected value. This created two very different amplitude approaches
between the noise measurement and RSSI measurement, which would
throw off the upper layer MS power control loop.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Previous removal of the energy detector requirement broke
the noise level calculation loop. The previous adaptive
approach was finicky - noticably at high gain levels. Since
we no longer use the energy threshold for primary burst gating,
we can return to a simpler world.
In the new approach, we compute a running average of energy
levels and track them with a noise vector. A timeslot that
passes the correlator threshold is a valid burst. These are
not used in the noise calculation. Everything else is
considered noise and used to compute the noise level with
respect to full scale input level, which for almost all
supported devices is 2^15.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Enabling the external reference on UHD devices through the configure
time switch is awkward. Use a database variable "TRX.Reference" with
'0' or '1' value for internal and external references respectively.
Use internal reference is no entry is defined.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
We want to push UHD logs to the OpenBTS logging system, but most
device errors occur at startup, so keep the output on stdout until
after device initialization. That way obvious errors are easily
viewable before seeing the useless TRX timeout message.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This primarily addresses the error case at initialization.
In the event that the transceiver fails to start, we should
be able cleanly shutdown and release while providing a useful
reason for exiting.
After the radio is started and threads launched, there
are no thread state variables or shutdown messaging between
threads, and the transceiver cannot be consistently
shutdown. This issue remains to be solved.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Previous send and receive buffers at the radio interface were
arbitrarily set to a sufficient size. For normal (non-resampling)
devices, use a block (chunk) size of 625 samples. For 64 or 100
MHz resampling devices, use 4 times the reduced resampling
numerator or denominator and provide bounds checking where
appropriate.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Move B100 to the resampling interface with default
clocking. This temporarily resolves undetermined
FPGA clocking issues. This also provides extensible
support for multiple clocking rates and resampling
ratios.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Current functionality with these old versions is questionable.
There is no reason to use any version of GNU Radio / libusrp older
than 3.3. Version 3.4.2 is the only recommended version for USRP1
users.
Non-USRP1 users must use UHD driver from Ettus Research.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Identical to B200 support, but explicitly check for the device type
name.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This patch applies oversampling, when selected with 4 sps,
to the downlink only, while running the receiver with
minimal sampling at 1 sps. These split sample rates allow
us to run a highly accurate downlink signal with minimal
distortion, while keeping receive path channel filtering
on the FPGA.
Without this patch, we oversample the receive path and
require a steep receive filter to get similar adjacent
channel suppression as the FPGA halfband / CIC filter
combination, which comes with a high computational cost.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This requires an additional memcpy() on the signal vector
constructor, but allows the interpolation filter to use
SSE optimzationed convolution.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This patch primarily addresses observed repeated overrun
conditions in embedded environments - namely ARM.
The heartbeat of the transceiver is derived from the receive
sample stream, which drives the main GSM clock. Detach the
transmit thread from the receive loop to avoid interfering with
the receive I/O, which is sensitive to overrun conditions if
pull process is interrupted.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Unsupported at 4 sps, and performance benefits remain
to be proven at 1 sps. Disable until further testing.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Start the correlation search window at 4 symbols before
the expected correlation peak. End the search at 10
symbols and 4 + maximum expected delay for RACH and TSC
bursts respectively.
This change lowers receive side cpu utilization while
maintaining reasonable timing jitter and accuracy tolerance.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
This patch only applies to resampling use at 4 samples-per-symbol.
By extention that means only USRP2 / N2xx devices are affected.
At 4 samples-per-symbol we restrict output bandwidth to roughly
roughly 700 MHz, which combined with the 2 pulse Laurent
approximation yields < 0.5 degrees of RMS phase error at the
resampler output.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Set master clock rate to 52 MHz for B200. Also, we want to avoid
floating point comparison errors on clock rate settings, but we
expect to be able really set the rates we specify. Set the
offset limit to 1 Hz. If we can't set our rates with that level
of precision, then something is wrong.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Provides substantially improved transmit phase error
performance when enabled. Requires use of 4 samples
per symbol, and is enabled by default when set.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
If there is an error in the sample rate determination, noted
by a negative return sample rate value, error directly and
don't try to set the device rate.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Both RACH and normal bursts are detected with the same approach of
midamble correlation combined with peak-to-average ratio. The
difference is the midamble placements and lengths. Thus, there is
no reason to have independent implementations.
This patch creates a common call burstDetect(), while leaving the
correlation window indexing in the original calls.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Requires Fairwaves UHD driver.
https://github.com/chemeris/UHD-Fairwaves.git
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
When 4 samples-per-symbol operation is selected, replace the
existing pulse approximation, which becomes inaccurate with
non-unit oversampling, with the primary pulse, C0, from the
Laurent linear pulse approximation.
Pierre Laurent, "Exact and Approximate Construction of Digital Phase
Modulations by Superposition of Amplitude Modulated Pulses", IEEE
Transactions of Communications, Vol. 34, No. 2, Feb 1986.
Octave pulse generation code for the first three pulses of the
linear approximation are included.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
The adaptive energy threshold gating suffers a near-far problem
at certain gain levels. This is due to exponential threshold
raising, but linear decreases. A large signal level followed by
a period low signal level causes (comparatively) weak signals to
go undetected. Additionally, the algorithm performs differently
at multiple RF gain levels.
This patch switches solely to correlation based gating for burst
detection. The main computational load with this approach is
sub-sample width peak interpolation, which we disable for intial
detection and run after threshold passing.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Convertions are performed in multiples of 4 or 8. All loads are
considered unaligned.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
|
Replace the polyphase filter and resampler with a separate
implementation using SSE enabled convolution. The USRP2 (including
derived devices N200, N210) are the only supported devices that
require sample rate conversion, so set the default resampling
parameters for the 100 MHz FPGA clock. This changes the previous
resampling ratios.
270.833 kHz -> 400 kHz (65 / 96)
270.833 kHz -> 390.625 kHz (52 / 75)
The new resampling factor uses a USRP resampling factor of 256
instead of 250. On the device, this allows two halfband filters to
be used rather than one. The end result is reduced distortial and
aliasing effecits from CIC filter rolloff.
B100 and USRP1 will no be supported at 400 ksps with these changes.
Signed-off-by: Thomas Tsou <tom@tsou.cc>
|
