• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 97 subscribers
  • Views 298 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Info on processRMsCmd

chris2476
Intellectual 310 points

Is there any info on what the command numbers mean for processRMsCmd?  I have not been able to find any in my DVSDK or on the TI website.  I know the CE logging task is part of RMS, but I have not been able to find much info on it.

Here is my specific issue: I am working with a DM6446, dvevm_1_20, codec_engine_1_10_01.  I am trying to restart (delete and create) a codec many times for stress testing.  It keeps failing on the 128th restart.  For successful restarts, I see a log like this:

[DSP] @0x78dbcd78:[T:0x8fb3b244] CR - processRmsCmd(0x8fe01928, 4056): cmd = 0
[DSP] @0x78dbcdbf:[T:0x8fb3b244] OM - Memory_alloc> Enter(size=0x18)
[DSP] @0x78dbcdf7:[T:0x8fb3b244] OM - Memory_alloc> return (0x8fb42b68)
[DSP] @0x78dbce30:[T:0x8fb3b244] OM - Memory_alloc> Enter(size=0xf)
[DSP] @0x78dbce62:[T:0x8fb3b244] OM - Memory_alloc> return (0x8fb42b80)
[DSP] @0x78dbce9d:[T:0x8fb3b244] OM - Memory_alloc> Enter(size=0x20)
[DSP] @0x78dbced0:[T:0x8fb3b244] OM - Memory_alloc> return (0x8fb42b90)
[DSP] @0x78dbcf0a:[T:0x8fb3b244] OM - Memory_alloc> Enter(size=0x24)
[DSP] @0x78dbcf3e:[T:0x8fb3b244] OM - Memory_alloc> return (0x8fb42bb0)
(followed by successful VISA_create call for the codec)

For the unsuccessful restart, I see this:

[DSP] @0x79cd6509:[T:0x8fb3b244] CR - processRmsCmd(0x8fe01928, 4056): cmd = 5
[DSP] @0x79cd6548:[T:0x8fb3b244] CR - remote time = 0x0, trace buffer size = 4032
(then the VISA_Create fails)

  • 0
    TI__Genius 17205 points

    I think this is the RMS_CmdSet enum, found in ti/sdo/ce/rms.h.  CE 1.10.01 didn't come with source code, so I've copied that enum, specifically from the 1.10.01 release, here:

    typedef enum RMS_CmdSet {
        RMS_CREATENODE,
        RMS_STARTNODE,
        RMS_DELETENODE,
        RMS_GETMEMSTAT,
        RMS_GETCPUSTAT,
        RMS_GETTRACE,
        RMS_GETVERS,
        RMS_WRITEWORD,
        RMS_SETTRACEMASK
    } RMS_CmdSet;

    Chris

  • 0 in reply to Chris Ring
    Intellectual 310 points

    Thanks for the info.

    So it looks like on the 128th try, for some reason the VISA_Create doesn't create a node and allocate buffers, but instead gets the trace data.

    I have been doing more testing.  The application deletes/creates the codec in one thread, and makes the process calls in a different thread.  A semaphore is used as a mutex so that only one thread at a time can access the codec.

    In the test that fails, the process task processes a few frames of video between restarts.

    If I do a tight loop and restart more than 128 times without letting the process task run (i.e. I hold on to the mutex in the delete/create task), then I do not see the problem.

    Any thoughts on this?  If there was a memory leak I should see the Memory_alloc calls fail, right?  It looks like they don't even happen.