Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
After fixing the race, this change uses Tom's idea to stop enqueuing
data when trying to stop, and adds a new flush_rx_samples method
to drop any samples that may have already been accumulated.
I ran Tom's test case 500 times with 0 failures ;-)
|
|
This fixes the bulk of the problem. Next step is to drop data packets
while waiting for the reply.
|
|
|
|
* Fixing a line in the clock recovery algorithm. This works with a bit larger error than there proably should be.
* Adding clock sync algorithm using PFB. This works, but needs a bit more work.
|
|
error than there proably should be.
|
|
* Move AC_PROG_CC into GR_FORTRAN macro definition
|
|
accidentally using the rx_clk in one stage of the fifos on the tx side.
|
|
* GR_FORTRAN macro leaves $ac_ext=f on some systems
* Restore by invoking AC_PROG_CC
Reported-by: Szymon Jakubczak <szym@csail.mit.edu>
Signed-off-by: Johnathan Corgan <jcorgan@corganenterprises.com>
|
|
|
|
|
|
"enhanced level logic" for accurate fullness. Maybe this will help...
|
|
|
|
|
|
|
|
and fifo_2clock.v are empty
|
|
|
|
|
|
|
|
|
|
|
|
* waterfall and fft use a common autoscale function
* Fix so that the waterfall texture is initialized with a buffer of the same size.
|
|
* Modifications to usrp2 source and sink so that set center freq is called afer set lo offset.
* Modifications to the usrp blocks and wrapper so that the lo offset is set with the lo frequency.
|
|
|
|
If the fft size was a non power of two, the buffer would be a different size.
This would cause a segfault. Particularly because fft_window was throwing out a bin.
|
|
afer set lo offset.
|
|
with the lo frequency.
|
|
* Don't call GR_SUBVERSION in configure.ac
* Remove --svn* options from gnuradio-config-info
|
|
* Modified log power fft block so ref scale is peak to peak.
* Tweaked fft sink autoscale routine to come up with better numbers.
* Modified scope sink ac couple block to use constant tap. The previous
tap calculation would cause failure for very small sample rates.
|
|
* Corresponds to SVN r11620:11622
* Fixes ticket 405
Reported-by: Dimitris Symeonidis <azimout@gmail.com>
Fixed-by: Matt Ettus <matt@ettus.com>
|
|
missing. MDIO seems ok.
|
|
|
|
Tweaked fft sink autoscale routine to come up with better numbers.
Modified scope sink ac couple block to use constant tap.
The previous tap calculation would cause failure for very small sample rates.
|
|
* svn diff http://gnuradio.org/svn/branches/developers/matt/new_eth
-r10782:11633
* Patch applied with no conflicts or fuzz.
|
|
|
|
* automatic store for empty traces on enable
* added traces to fft window
* added quadradio method to enable/disable cal
* use enum for band select
* added get band select
|
|
* Add shorthand for making tuples.
* Add blobs and shorthand pmt pseudo-constructors.
* gruel::send can now send to a pmt.
* Add top-level msg_passing.h include file.
* QA code now works.
* Add pmt type that wraps a gruel::msg_accepter.
|
|
Added files to Makefile.am for distribution.
Added colorbar to spectrogram for magnitude measurement.
Fixed spectrogram plotting axis.
Adding a spectrogram plot. The axis need work.
GR plotter tool handles end of file and files shorter than the block length.
Manages window/filter types better. Disables filter types if the designing algorithm doesn't support it (no RRC or Gaussian for equiripple filters).
User messages if PyQt and PyQwt are not installed (or found).
Setting validators for all edit boxes.
Fixing tab order.
Adding display for the number of taps in the filter.
Adding equiripple band reject filter to filter design app.
Adding P-M version of band reject filter.
Fixed documentation for optfir band pass filters.
Added routine for optfir equiripple filter design code to create complex bandpass filters. Also adds this ability to the filter designer.
Added design for Guassian filters.
Added design for RRC filters.
Added Band Notch filter
Adding complex bandpass filter design (for windowed filters only).
Adding plotting of group delay.
Adding labels to the plots.
...
|
|
|
|
|
|
|
|
|
|
|
|
algorithm doesn't support it (no RRC or Gaussian for equiripple filters).
|
|
|
|
|
|
|
|
|