|
I have taken an existing example (codec_bios) which just performs simple mic / speaker loopback on a codec interrupt. I added 3 lines of C: #include log.h #inlcude codec_bioscfg.h //and in main LOG_printf(&mylog, "starting"); I changed the DSP/BIOS Config so that RTDX is enabled, and the Host Channel Manager link type to RTDX (so channels are allowed). This basically loads the relevant RTDX libraries on the target along with the main program. I also added a log trace called "mylog" for the logging. Basically the changes I made were minimal to make the RTDX LOG functionality work. When the program runs and the Enable RTDX checkbox is checked in CCS, regular glitches are seen on the oscilloscope (looking at speaker o/p) in the sine wave. When Enable RTDX is unchecked, its perfect - no glitches. If the rate at which the PC polls the board is increased, the glitches happen more often, therefore it can probably be deduced that the target is interrupted when the PC polls the target. Even though TI say that RTDX works "without interfering with the target application"! I am thinking other people must have come accross this problem, as it happens in several different programs! Thanks in advance for any help, Andrew Edwards |
|
|
LOG_printf is affecting real time!
Started by ●April 8, 2002
Reply by ●April 8, 20022002-04-08
|
On Mon, 08 Apr 2002 11:55:14 -0000, you wrote: >If the rate at which the PC polls the board is increased, the >glitches happen more often, therefore it can probably be deduced that >the target is interrupted when the PC polls the target. Even though >TI say that RTDX works "without interfering with the target >application"! > This is from page 71 of spru423.pdf (DSP/BIOS user manual): LOG_event and LOG_printf both operate on logs with interrupts disabled. This allows hardware interrupts and other threads of different priorities to write to the same log without having to worry about synchronization. LOG_ is supposed to take about 32 cycles, but if you have multiple threads calling it you may be punching holes in the audio with them. Try playing with turning various logging/statistics options on and off to see if one of them effect things more than the other. Brian ----------------- Brian C. Lane (KC7TYU) Programmer www.shinemicro.com RF, DSP & Microcontroller Design |
