Other Parts Discussed in Thread: AM5728
Tool/software: Code Composer Studio
When attempting to use the gdb_agent_console from CCS 7.4 with TI Emulators 7.0.100.0 under Windows 10 with a board data file generated by a CCS test data connection for an AM5728 connected to a Blackhawk USB560-M, initially attempted to give a relative path to the board data file. However, using a relative path causes a "No CPUs that support GDB are defined in configuration file" error to be reported:
C:\Users\mr_halfword\E2E_example_projects\AM5728_GCC_class_variables>c:\ti\ccs740\ccsv7\ccs_base\common\uscif\gdb_agent_console.exe targetConfigs\testBoard_bh560usbm.dat No CPUs that support GDB are defined in configuration file
If an absolute path to the same target data file is given then the gdb_agent_console initializes without error:
C:\Users\mr_halfword\E2E_example_projects\AM5728_GCC_class_variables>c:\ti\ccs740\ccsv7\ccs_base\common\uscif\gdb_agent_console.exe c:\Users\mr_halfword\E2E_example_projects\AM5728_GCC_class_variables\targetConfigs\testBoard_bh560usbm.dat CPU Name Port -------- ---- c66xx_dsp1 :55000 c66xx_dsp2 :55001 cortex_m4_ipu1_c0 :55002 cortex_m4_ipu1_c1 :55003 cortex_m4_ipu2_c0 :55004 cortex_m4_ipu2_c1 :55005 cortexa15_0 :55006 cortexa15_1 :55007 Starting all cores CPU Name Status -------- ------ cortexa15_1 Waiting for client c66xx_dsp1 Waiting for client c66xx_dsp2 Waiting for client cortex_m4_ipu1_c0 Waiting for client cortexa15_0 Waiting for client cortex_m4_ipu2_c1 Waiting for client cortex_m4_ipu2_c0 Waiting for client cortex_m4_ipu1_c1 Waiting for client
When the SysInternals Process Monitor was used to trace the file I/O performed by gdb_agent_console, in the failure case with a relative board data file then gdb_agent_console made some attempts to open C:\ti\ccs740\ccsv7\ccs_base\common\uscif\targetConfigs\testBoard_bh560usbm.dat, which failed with PATH NOT FOUND.
Using Processor Explorer to look at the working gdb_agent_console shows that the currently directory is set to c:\ti\ccs740\ccsv7\ccs_base\common\uscif\. I.e. the problem seems to be that at some point gdb_agent_console changes the working directory of the process away from that where it was started, leading to a failure to open a relative board data file given on the command line.
I think this may be the same issue for which DBGTRC-3927 was raised in the original thread, and a work-around for the defect could be to always specify an absolute path for the board data file.
The attached zip file contains the board data file used, and the Process Monitor capture for the failure cause with a relative board data file path. ccs740_gdb_agent_relative_path.zip
