Shared memory for IPC buffers/Concerto C28

Sabina,

Screen shots would help with the errors and ROV details.  Its hard to parse the ROV details.  It does look like they are correctly created in the right memory region.

How are you allocating your buffers, what's the call you are making look like?  You should be able to specify the heapBuf that you created.

Judah

Hi Judan,

I have the following error in the beginning of the program running:
[C28xx_0] xdc.runtime.Memory: line 52: out of memory: heap=0x9b92, size=42
[C28xx_0] MessageQ_alloc PMCU type failed

Pls see our defenitions in the attached file: 5822.L_C28.rar

 In the code:

for MsgQHeapBuf0:
typedef struct {
 MessageQ_MsgHeader  header;   // The MessageQ header
 Uint16    MsgType;  // The message type
 Uint16    Data[100];  // Message data
} MY_M3_C28_Msg;

for PMCUMsgQHeapBuf1:
 typedef struct {
 MessageQ_MsgHeader  header;   // The MessageQ header
 Uint16    MsgType;  // The message type
 Uint16    Data[25];  // Message data
} PMCU_TYPE_M3_C28_Msg;

void IPC_init_1(void)
{
 Int  status;
    UInt numProcs = MultiProc_getNumProcessors(); // Number of processors, in the Concerto case = 2

    // Generate queue names based on own proc ID and total number of procs
    System_sprintf(localQueueName, "%s_queue",
            MultiProc_getName(MultiProc_self()));
    System_sprintf(remoteQueueName, "%s_queue",
            MultiProc_getName(1 - MultiProc_self()));

    // Determine which processors Notify will communicate with based on the local MultiProc id.
    srcProc = ((MultiProc_self() - 1 + numProcs) % numProcs);
    dstProc = ((MultiProc_self() + 1) % numProcs);

    // Register call back with Notify. It will be called when the processor
    // with id = srcProc sends event number EVENTID to this processor.
    status = Notify_registerEvent(srcProc, INTERRUPT_LINE, EVENTID,
                                  (Notify_FnNotifyCbck)NotifycallBackFxn, NULL);
    if (status < 0) {
        System_abort("Notify_registerEvent failed\n");
    }

    status = Notify_registerEvent(srcProc, INTERRUPT_LINE, EVENTID_ERR,
                                      (Notify_FnNotifyCbck)NotifycallBackFxn_ERR, NULL);
 if (status < 0) {
  System_abort("Notify_registerEvent failed\n");
 }

 status = Notify_registerEvent(srcProc, INTERRUPT_LINE, EVENTID_IPC,
                                       (Notify_FnNotifyCbck)NotifycallBackFxn_IPC, NULL);
  if (status < 0) {
   System_abort("Notify_registerEvent failed\n");
  }

    // Register default system heap with MessageQ
    MessageQ_registerHeap((IHeap_Handle)(MsgQHeapBuf0),HEAPID);

    // Create the local message queue
    messageQ = MessageQ_create(localQueueName, NULL);
    if (messageQ == NULL) {
        System_abort("MessageQ_create failed\n" );
    }

    // Open the remote message queue. Spin until it is ready.
    do {
        status = MessageQ_open(remoteQueueName, &remoteQueueId);
        // Sleep for 1 clock tick to avoid inundating remote processor
        // with interrupts if open failed
        if (status < 0) {
            Task_sleep(1);
        }
    } while (status < 0);

}

/*******************************************************************************
Function Name : IPC_init_2
Description :   This function init's the IPC data, for messageQ2
*******************************************************************************/
void IPC_init_2 (void)
{
 Int  status;
    UInt numProcs = MultiProc_getNumProcessors(); // Number of processors, in the Concerto case = 2

    // Generate queue names based on own proc ID and total number of procs
    System_sprintf(localQueueName_tsk2, "%s2_queue",
            MultiProc_getName(MultiProc_self()));
    System_sprintf(remoteQueueName_tsk2, "%s2_queue",
            MultiProc_getName(1 - MultiProc_self()));

    // Register default system heap with MessageQ
    MessageQ_registerHeap((IHeap_Handle)(PMCUMsgQHeapBuf1), HEAPID_2);

    // Create the local message queue
    messageQ_2 = MessageQ_create(localQueueName_tsk2, NULL);
    if (messageQ_2 == NULL) {
        System_abort("MessageQ2_create failed\n" );
    }

    // Open the remote message queue. Spin until it is ready.
    do {
        status = MessageQ_open(remoteQueueName_tsk2, &remoteQueueId_2);
        // Sleep for 1 clock tick to avoid inundating remote processor
        // with interrupts if open failed
        if (status < 0) {
            Task_sleep(1);
        }
    } while (status < 0);

}

/*******************************************************************************
Function Name : IPC_allocMsg
Description :   Allocate the IPC message data,
*******************************************************************************/
MY_M3_C28_Msg * IPC_allocMsg ( void )
{
 MY_M3_C28_Msg *msgAlloc;

    // Allocate a message
 msgAlloc = (MY_M3_C28_Msg *)MessageQ_alloc(HEAPID, sizeof(MY_M3_C28_Msg));
    if (msgAlloc == NULL) {
       System_abort("MessageQ_alloc failed\n" );
    }
    return(msgAlloc);
}

/*******************************************************************************
Function Name : IPC_allocMsgPMCUType
Description :   Allocate the IPC message data,
*******************************************************************************/
PMCU_TYPE_M3_C28_Msg * IPC_allocMsgPMCUType ( void )
{

 PMCU_TYPE_M3_C28_Msg *msgAlloc;

 // Allocate a message
 msgAlloc = (PMCU_TYPE_M3_C28_Msg *)MessageQ_alloc(HEAPID_2, sizeof(PMCU_TYPE_M3_C28_Msg));
 if (msgAlloc == NULL) {
    System_abort("MessageQ_alloc PMCU type failed\n" );
 }
 return(msgAlloc);
}  

It seems that we have some problems in the buffers definitions into  the shared memory. Without shared memory using it run properly.

Thanks, Sabina

Sabina,

First, I think I understand why you are getting the ROV error.  In your .cfg file, you have not placed the "IPC_BUFFERS_SECTION".  You've created the section by the following code, but the section did not get placed and therefore gets assigned to address starting at 0x0 by default.

    var heapBuf0Params = new HeapBuf.Params();
    heapBuf0Params.instance.name = "MsgQHeapBuf0";
    heapBuf0Params.blockSize = 256;
    heapBuf0Params.numBlocks = 4;
    heapBuf0Params.align = 8;
    heapBuf0Params.sectionName = "IPC_BUFFERS_SECTION";
    Program.global.MsgQHeapBuf0 = HeapBuf.create(heapBuf0Params);

What you need is the following additional line.  This will place the output section "IPC_BUFFERS_SECTION" to the memory segment "IPC_BUFFERS_SECTION".  I noticed this by looking at your .map file.  It shows your IPC_BUFFERS_SECTION segment is unused and shows the section IPC_BUFFERS_SECTION starts at 0x0.

    Program.sectMap["IPC_BUFFERS_SECTION"] = { loadSegment: "IPC_BUFFERS_SECTION" };

Now don't know if this will resolve the rest of the issue.  I will continue looking at that.

Judah

Sabina,

Looks like you are running out of heap blocks on the HEAPID_2.  Are you able to monitor using ROV, the number of allocs and frees?

I assume this probably doesn't fail right away since you have 8 blocks according to your .cfg file?  Who's calling free on the alloc'ed buffers?

I didn't get your last statement about buffer definition into the shared memory.

Judah

Hi Judah,

The 1st point about configuration file:
I put in the configuration file section IPC_BUFFERS_SECTION. Pls see line#91 and #157 in message_c28.cfg that I sent in L_C28.rar. But I really don't see it in the MAP file.
May be we have any conflict between M3 and C28 configuration?
I send you our M3 cfg , cmd and map 2022.L_M3.rar

The 2nd point about HEAPID_2.
We don't see any overflow in ROV during application working in the case if we don't use shared memory for IPC buffers
we define shared memory in m3:
    RAMMReqSharedMemAccess(S0_ACCESS , SX_C28MASTER );
    RAMMReqSharedMemAccess(S1_ACCESS , SX_C28MASTER );
    RAMMReqSharedMemAccess(S2_ACCESS , SX_C28MASTER );
    RAMMReqSharedMemAccess(S3_ACCESS , SX_C28MASTER );

Our code for allocation and free: (two tasks)

Void Tsk_IPCMsgRecieveFromLPU(UArg arg0, UArg arg1)
{
 MY_M3_C28_Msg  *myRcvMessage;
    // Init the IPC, M3 <-> C28x communication for MessageQ1  (LPU messages)
 IPC_init_1 ();
 System_printf("CIPCInterface (C28)::before while after IPC_init_1() \n");
    while (1)
    {
     // receive a message
     myRcvMessage = IPC_recMsgQ1 ();
     MsgType=myRcvMessage->MsgType;
     memcpy(pDataBuff,  myRcvMessage->Data, 100);
     IPC_freeMsg (myRcvMessage);
     (((CBoardSubSystem*)g_pBoard)->m_pLPUInterface)->LPUMessageHandle(MsgType, pDataBuff);
      }

}

Void Tsk_IPCMsgRecieveFromPMCU (UArg arg0, UArg arg1)
{
  PMCU_TYPE_M3_C28_Msg *myRcvMessage2;
    // Init the IPC, M3 <-> C28x communication  for MessageQ2  (PMCU messages)
 IPC_init_2 ();
 System_printf("CIPCInterface (C28):: before while after IPC_init_2() \n");

    while (1)
    {
     // receive a message
     myRcvMessage2 = IPC_recMsgQ2 ();
     (((CBoardSubSystem*)g_pBoard)->m_pPMCUInterface)->PMCUMessageHandle(myRcvMessage2->MsgType, myRcvMessage2->Data);
     //Free the received message
     IPC_freeMsgPMCUType (myRcvMessage2);
    }

}

Thanks, Sabina

Sabina,

I don't think you understood what I said.  I looked at your c28 files that you attached.  Doing line #91 and #157 is not enough.  Those two lines simply put those two HeapBufs in a linker output section called "IPC_BUFFERS_SECTION".

But don't you want to put this output section into the memory segment "IPC_BUFFERS_SECTION"?  The confusing part is you have both an output section called "IPC_BUFFERS_SECTION" and a memory segment called "IPC_BUFFERS_SECTION".

You need the additional line that I posted in order to actually place your output sectoin "IPC_BUFFERS_SECTION" into the memory segment named "IPC_BUFFERS_SECTION".

Judah

Sabina,

I looked at your M3 *.cfg file.  You have not specified who has write access to which shared memory segment.  Is the C28 supposed to have write access to some shared memory?  You need to specify this via the following config parameter.  Please see the IpcMgr.xdc file for more details.

   IpcMgr.sharedMemoryOwnerMask

Judah

Hi Judah,

Now the situation is much better! I seems that programm work properly.
I added per your recomendation two definitions:
1. in C28.cfg :  Program.sectMap["IPC_BUFFERS_SECTION"] = { loadSegment: "IPC_BUFFERS_SECTION" };
2. in M3.cfg
IpcMgr.readAddr  = 0x20008000; //sram0
IpcMgr.readAddr  = 0x2000A000; //sram1
IpcMgr.readAddr  = 0x2000C000; //sram2
IpcMgr.readAddr  = 0x2000E000; //sram3
IpcMgr.sharedMemoryOwnerMask = 0x0F;

Attached is my updated cfg,cmd and map files
0638.L_M3_1.rar
6648.L_C28_1.rar

Could you help us to resolve one additional problem in this thread (connected to IPC also)?
I get some warning that can not understand:

 warning #339-D: linkage specification is incompatible with previous "MsgQHeapBuf0" (declared at line 23 of "../Debug/configPkg/package/cfg/BlueBox_pem3.h") ‪.xdchelp‬ ‪/L-Full_M3-1‬ 42 C/C++ Problem
 warning #339-D: linkage specification is incompatible with previous "NotifySemaphore0" (declared at line 32 of "../Debug/configPkg/package/cfg/BlueBox_pem3.h") ‪.xdchelp‬ ‪/L-Full_M3-1‬ 44 C/C++ Problem
 warning #339-D: linkage specification is incompatible with previous "PMCUMsgQHeapBuf1" (declared at line 77 of "../Debug/configPkg/package/cfg/BlueBox_pem3.h") ‪.xdchelp‬ ‪/L-Full_M3-1‬ 43 C/C++ Problem

line 23 in BlueBox_pem3.h:  extern const ti_sysbios_heaps_HeapBuf_Handle MsgQHeapBuf0;
line 32 in BlueBox_pem3.h: extern const ti_sysbios_knl_Semaphore_Handle NotifySemaphore0;
line 77 in BlueBox_pem3.h: extern const ti_sysbios_heaps_HeapBuf_Handle PMCUMsgQHeapBuf1;

The following pacage include this file if you need it.
4606.M3configPkg.rar

What could be the reason of these warnings?

Thanks for your help, Sabina

Sabina,

I don't know what's causing these warnings.  Are the warnings coming from the compiler or the linker?

In your 'c' files that use these objects, are you declaring these objects a different way or are you simply just including the header file?

Judah

Hi Judan,

It is linkage warnings. the same warnings we get for c28 and m3

in our c file we use include

#include <xdc/runtime/System.h>
#include <xdc/runtime/Memory.h>
#include <xdc/runtime/IHeap.h>

and don't declare objects at all.

Thanks,Sabina

Sabina,

Are you sure its the linker not the compiler giving this warning?  The warning look like a compile warning to me.

Can you add this #include after all your other #includes into every .c file that is using one of those global objects from your .cfg file:

#include <xdc/cfg/global.h>

Judah

Hi Judan,

The problem resolved, we had an extern in ipc.h file. now we removed these extern and added <xdc/cfg/global.h> per ypur recomendation. thanks!

but now we have another problem
during running program without gel file we get the following error:

C28xx_0: Loader: One or more sections of your program falls into a memory region that is not writable.  These regions will not actually be written to the target.  Check your linker configuration and/or memory map.
i didn't find any conflict in map file (sent you in the past it in the attached). could you see our map and cfg again.

Thanks for your help, Sabina

Hi Judan,

Pls help us to resolve one last problem that we have in this post. all work fine from the CCS debugger. But we still have problem with shared memory after starting program from the flash without debugging from CCS.. IPC doesn't work in this case - it seems that
MessageQ_get(messageQ, &msg, MessageQ_FOREVER); and MessageQ_get(messageQ_2, &msg, MessageQ_FOREVER);
don't returned with new message at all.
without shared memory using we don't have this problem.

Where could be the problem here?  

Thanks, Sabina

Sabina,

What is your gel file doing?  Is this a gel file that you authored or is it authored by TI? Can you attach the gel file here?

Seems like the gel file must be initializing some memory which are you loading to and without this, the memory is not writable.

This could also explain why when you run from flash its not working.  Sounds to me like whatever the gel file is doing needs to be done at runtime.

Judah

Hi Judah,

We use standard gel file from TI. Only tomorrow I could send it to you.   

Thanks, Sabina

our gel files attached:
2210.JelFiles.rar

Sabina

hi Judan,

Could you answer us and help to finish this issue.

Thanks,Sabina  

H Judan,

it didn't help.

So i close this post and reopen another one in c2000 forum

Thanks for your help, Sabina