Memory leak in SPI transfer

[KirkAtWork]

In the past there was a memory leak in the SPI0.read function. It was fixed.

Now I'm seeing a problem with the SPI0.transfer function. The Python code below demonstrates the problem. The amount of memory "leaked" is proportional to the number of words transferred. Un-comment one of the lines in the for loop to observe the problem. Sample output is included below also.

Thanks very much for taking a look, Kirk Mitchell Kollmorgen

****\ python script to demonstrate the memory leak

pybbioMemoryLeakV102.py

PyBBIO SPI memory leak demonstration

from bbio import * import resource

initialize the SPI to talk to the ADXL312 accelerometer chip

def setup(): SPI0.begin() SPI0.setClockMode(0,3) SPI0.setMaxFrequency(0, 5000000) SPI0.setBitsPerWord(0, 16)

read blocks of words from the accelerometer

def loop(): for j in range(10): print "***Reading 100000 samples, block=",j mem_used = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss print 'Memory usage( kB)=', mem_used

    for i in range(100000):
        #The following 3 lines of code demonstrate the memory leak (use only 1 at a time)

        #spiReadData = SPI0.read(0, 1)  #this line does not leak
        #spiReadData = SPI0.transfer(0, [0, 0, 0, 0]) #this line leaks
        spiReadData = SPI0.transfer(0, [0, 0, 0, 0, 0, 0, 0, 0])    #this line leaks twice as fast as the line above

    delay( 1000 )

print "Cleaning up & exiting"
SPI0.end
stop()

run(setup, loop)

*** Output from 2 runs of the above code follows ****

root@beaglebone:/kirkpy# python pybbioMemoryLeakV102.py **Reading 100000 samples, block= 0 Memory usage( kB)= 4372 ***Reading 100000 samples, block= 1 Memory usage( kB)= 10552 ****_Reading 100000 samples, block= 2 Memory usage( kB)= 15280 _Reading 100000 samples, block= 3 Memory usage( kB)= 20008 ****_Reading 100000 samples, block= 4 Memory usage( kB)= 24732 _Reading 100000 samples, block= 5 Memory usage( kB)= 29456 ****_Reading 100000 samples, block= 6 Memory usage( kB)= 34184 _Reading 100000 samples, block= 7 Memory usage( kB)= 38908 ****_Reading 100000 samples, block= 8 Memory usage( kB)= 43636 _***Reading 100000 samples, block= 9 Memory usage( kB)= 48360 Cleaning up & exiting

root@beaglebone:/kirkpy# python pybbioMemoryLeakV102.py **Reading 100000 samples, block= 0 Memory usage( kB)= 4372 ***Reading 100000 samples, block= 1 Memory usage( kB)= 15276 ****_Reading 100000 samples, block= 2 Memory usage( kB)= 24732 _Reading 100000 samples, block= 3 Memory usage( kB)= 34184 ****_Reading 100000 samples, block= 4 Memory usage( kB)= 43636 _Reading 100000 samples, block= 5 Memory usage( kB)= 53084 ****_Reading 100000 samples, block= 6 Memory usage( kB)= 62536 _Reading 100000 samples, block= 7 Memory usage( kB)= 71988 ****_Reading 100000 samples, block= 8 Memory usage( kB)= 81440 _***Reading 100000 samples, block= 9 Memory usage( kB)= 90892 Cleaning up & exiting

[alexanderhiam]

Thanks for another good catch, Kirk! It's fixed now on Github and I'll push another release to PyPI soon.

[KirkAtWork]

Sorry about the weird bold text in the post. Not sure what happened. I really appreciate your quick response to this. I'll test it and let you know how it goes. Thanks again, Kirk