linuxgnuradiousrpuhd

flowgraph fails with "LLLL..." for one network adapter, succeeds with other adapter


I want to execute the usrp_echotimer_dual_cw example from the GNURadio gr-radar OOT module.

The flowgraph works fine with the internal gigabit ethernet adapter but fails with the external PCI gigabit ethernet adapter.

Here is the output of the successful execution (eth0) of the flowgraph:

Executing: "/home/christophe/new/examples/usrp/usrp_echotimer_dual_cw.py"

linux; GNU C++ version 4.8.4; Boost_105400; UHD_003.010.git-101-g4a1cb1f2

Using Volk machine: avx_64_mmx_orc
-- Opening a USRP2/N-Series device...
-- Current recv frame size: 1472 bytes
-- Current send frame size: 1472 bytes

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The send buffer could not be resized sufficiently.
    Target sock buff size: 1048576 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.wmem_max=1048576
Using USRP Device (TX): 
Single USRP:
  Device: USRP2 / N-Series Device
  Mboard 0: N210r4
  RX Channel: 0
    RX DSP: 0
    RX Dboard: A
    RX Subdev: RFX2400 RX
  TX Channel: 0
    TX DSP: 0
    TX Dboard: A
    TX Subdev: RFX2400 TX

Setting TX Rate: 14250000

UHD Warning:
    The requested interpolation is odd; the user should expect CIC rolloff.
    Select an even interpolation to ensure that a halfband filter is enabled.
    interpolation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.250000 MHz)

UHD Warning:
    The hardware does not support the requested TX sample rate:
    Target sample rate: 14.250000 MSps
    Actual sample rate: 14.285714 MSps
Actual TX Rate: 1.42857e+07

UHD Warning:
    The requested interpolation is odd; the user should expect CIC rolloff.
    Select an even interpolation to ensure that a halfband filter is enabled.
    interpolation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.285714 MHz)
-- Opening a USRP2/N-Series device...
-- Current recv frame size: 1472 bytes
-- Current send frame size: 1472 bytes

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The send buffer could not be resized sufficiently.
    Target sock buff size: 1048576 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.wmem_max=1048576
Using USRP Device (RX): 
Single USRP:
  Device: USRP2 / N-Series Device
  Mboard 0: USRP2 r3
  RX Channel: 0
    RX DSP: 0
    RX Dboard: A
    RX Subdev: RFX2400 RX
  TX Channel: 0
    TX DSP: 0
    TX Dboard: A
    TX Subdev: RFX2400 TX

Setting RX Rate: 14250000

UHD Warning:
    The requested decimation is odd; the user should expect CIC rolloff.
    Select an even decimation to ensure that a halfband filter is enabled.
    decimation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.250000 MHz)

UHD Warning:
    The hardware does not support the requested RX sample rate:
    Target sample rate: 14.250000 MSps
    Actual sample rate: 14.285714 MSps
Actual RX Rate: 1.42857e+07

UHD Warning:
    The requested decimation is odd; the user should expect CIC rolloff.
    Select an even decimation to ensure that a halfband filter is enabled.
    decimation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.285714 MHz)
set_min_output_buffer on block 5 to 4194304
set_min_output_buffer on block 6 to 4194304
set_min_output_buffer on block 7 to 4194304
set_min_output_buffer on block 8 to 4194304
set_min_output_buffer on block 9 to 4194304
set_min_output_buffer on block 17 to 4194304
set_min_output_buffer on block 18 to 4194304
set_min_output_buffer on block 19 to 4194304
set_min_output_buffer on block 20 to 4194304
set_min_output_buffer on block 21 to 4194304
set_min_output_buffer on block 22 to 4194304
// Print results
rx_time: 2:0.0476682 
velocity: -5.82422 
range: 4.83493 

// Print results
rx_time: 2:0.411017 
velocity: 0.416016 
range: 4.80168 

// Print results
rx_time: 2:0.770647 
velocity: 1.24805 
range: 4.66541 

// Print results
rx_time: 3:0.122836 
velocity: 1.24805 
range: 4.86308 

// Print results
rx_time: 3:0.489277 
velocity: 1.66406 
range: 4.80136 

// Print results
rx_time: 3:0.848081 
velocity: -15.8086 
range: 5.37644 

// Print results
rx_time: 4:0.198079 
velocity: 4.57617 
range: 5.10404 

// Print results
rx_time: 4:0.558055 
velocity: 4.99219 
range: 4.4827 

// Print results
rx_time: 4:0.917475 
velocity: -0.832031 
range: 4.62831 

// Print results
rx_time: 5:0.266247 
velocity: 6.65625 
range: 5.24577 

// Print results
rx_time: 5:0.625892 
velocity: -6.65625 
range: 5.5386 

The failed execution (eth1) looks like this:

Executing: "/home/christophe/new/examples/usrp/usrp_echotimer_dual_cw.py"

linux; GNU C++ version 4.8.4; Boost_105400; UHD_003.010.git-101-g4a1cb1f2

Using Volk machine: avx_64_mmx_orc
-- Opening a USRP2/N-Series device...
-- Current recv frame size: 1472 bytes
-- Current send frame size: 1472 bytes

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The send buffer could not be resized sufficiently.
    Target sock buff size: 1048576 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.wmem_max=1048576
Using USRP Device (TX): 
Single USRP:
  Device: USRP2 / N-Series Device
  Mboard 0: N210r4
  RX Channel: 0
    RX DSP: 0
    RX Dboard: A
    RX Subdev: RFX2400 RX
  TX Channel: 0
    TX DSP: 0
    TX Dboard: A
    TX Subdev: RFX2400 TX

Setting TX Rate: 14250000

UHD Warning:
    The requested interpolation is odd; the user should expect CIC rolloff.
    Select an even interpolation to ensure that a halfband filter is enabled.
    interpolation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.250000 MHz)

UHD Warning:
    The hardware does not support the requested TX sample rate:
    Target sample rate: 14.250000 MSps
    Actual sample rate: 14.285714 MSps
Actual TX Rate: 1.42857e+07

UHD Warning:
    The requested interpolation is odd; the user should expect CIC rolloff.
    Select an even interpolation to ensure that a halfband filter is enabled.
    interpolation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.285714 MHz)
-- Opening a USRP2/N-Series device...
-- Current recv frame size: 1472 bytes
-- Current send frame size: 1472 bytes

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The recv buffer could not be resized sufficiently.
    Target sock buff size: 50000000 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.rmem_max=50000000

UHD Warning:
    The send buffer could not be resized sufficiently.
    Target sock buff size: 1048576 bytes.
    Actual sock buff size: 1000000 bytes.
    See the transport application notes on buffer resizing.
    Please run: sudo sysctl -w net.core.wmem_max=1048576
Using USRP Device (RX): 
Single USRP:
  Device: USRP2 / N-Series Device
  Mboard 0: USRP2 r3
  RX Channel: 0
    RX DSP: 0
    RX Dboard: A
    RX Subdev: RFX2400 RX
  TX Channel: 0
    TX DSP: 0
    TX Dboard: A
    TX Subdev: RFX2400 TX

Setting RX Rate: 14250000

UHD Warning:
    The requested decimation is odd; the user should expect CIC rolloff.
    Select an even decimation to ensure that a halfband filter is enabled.
    decimation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.250000 MHz)

UHD Warning:
    The hardware does not support the requested RX sample rate:
    Target sample rate: 14.250000 MSps
    Actual sample rate: 14.285714 MSps
Actual RX Rate: 1.42857e+07

UHD Warning:
    The requested decimation is odd; the user should expect CIC rolloff.
    Select an even decimation to ensure that a halfband filter is enabled.
    decimation = dsp_rate/samp_rate -> 7 = (100.000000 MHz)/(14.285714 MHz)
set_min_output_buffer on block 5 to 4194304
set_min_output_buffer on block 6 to 4194304
set_min_output_buffer on block 7 to 4194304
set_min_output_buffer on block 8 to 4194304
set_min_output_buffer on block 9 to 4194304
set_min_output_buffer on block 17 to 4194304
set_min_output_buffer on block 18 to 4194304
set_min_output_buffer on block 19 to 4194304
set_min_output_buffer on block 20 to 4194304
set_min_output_buffer on block 21 to 4194304
set_min_output_buffer on block 22 to 4194304
DReceive timeout before all samples received...
ULLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLterminate called after throwing an instance of 'std::runtime_error'
  what():  Receiver error ERROR_CODE_OVERFLOW (Out of sequence error)

These are my network cards:

[0] christophe:~ % lspci -nn | grep -i ethernet         
00:19.0 Ethernet controller [0200]: Intel Corporation 82579LM Gigabit Network Connection [8086:1502] (rev 04)
02:05.0 Ethernet controller [0200]: Intel Corporation 82541PI Gigabit Ethernet Controller [8086:107c] (rev 05)

I don't think it has to do with the performance of the computer, because that flowgraph works fine on a much weaker laptop pc (intel core 2 duo), but fails on my i7 desktop pc.


Solution

  • 00:19.0 Ethernet controller [0200]: Intel Corporation 82579LM Gigabit Network Connection [8086:1502] (rev 04)
    

    The 82579LM is the single known non-USB gigabit ethernet adapter that randomly drops packets without giving notice to the Operating System.

    I'm afraid you will have to use a different PCIe-Gigabit adapter.

    By the way, L means that your packet (containing some command for the USRP to execute, normally) is later than the time that was specified for the command to happen.

    The source code where the L get's printed:

    else if (metadata.event_code &
        async_metadata_t::EVENT_CODE_TIME_ERROR
    ) UHD_MSG(fastpath) << "L";
    

    Now, EVENT_CODE_TIME_ERROR means that

    Packet had time that was late.

    Which might be an effect of your network card dropping packets earlier, so that somehow the timing commands got out of order, or your application got confused. Normally, it's an indication of an application misdesign, but as your network hardware definitely is buggy (seen that controller fail many times, sorry), I would fix that first before investigating further.