view status.cpp @ 400:8c7453fb5bd9 api-inversion

Invert audioDB::power_flag / audiodb_power() Here the exciting discovery is that the mmap(), memcpy(), munmap() sequence is in fact not safe. In principle an msync() call should be inserted before unmapping for in-core changes to mmap()ed files to be flushed to disk. In this case we work around the problem entirely, by not mmap()ing anything and doing everything with file descriptors. Amusingly, that's probably not desperately safe either, this time because we have to move the file descriptor position (which is also a shared resource). dup() doesn't save us, as the duplicate file descriptor shares a file position. This applies also to the filling of data_buffer in the query loop, and in fact basically any call to lseek(), which is why I'm not fixing it now. Solution: if you have multiple threads all acting at once on a single database, do one audiodb_open() per thread, for now at least.
author mas01cr
date Thu, 27 Nov 2008 16:22:52 +0000
parents a65b31660804
children 58b88ab69424
line wrap: on
line source
#include "audioDB.h"
extern "C" {
#include "audioDB_API.h"
}

int audiodb_status(adb_t *adb, adb_status_t *status) {
  /* FIXME: it would be nice to be able to test for "is this database
     pointer valid", but at the moment we punt that to memory
     discipline.  */

  unsigned dudCount = 0;
  unsigned nullCount = 0;

  size_t trackTableLength = ALIGN_PAGE_UP(adb->header->numFiles * O2_TRACKTABLE_ENTRY_SIZE);
  unsigned *trackTable = 0;
  void *tmp = 0;
  if (adb->header->length > 0)  {
    tmp = mmap(0, trackTableLength, PROT_READ, MAP_SHARED, adb->fd, adb->header->trackTableOffset);
    if (tmp == (void *) -1) {
      return 1;
    }
    trackTable = (unsigned *) tmp;
  }

  for(unsigned k = 0; k < adb->header->numFiles; k++) {
    /* FIXME: this bare "16" here reveals a problem (or maybe two).
     * 16 here means the default value of the sequenceLength parameter
     * initializer (both in C++ and corresponding to the "-l" or
     * "--sequencelength" command-line argument).
     *
     * The problem is that the API as currently designed provides no
     * way to pass that information in to this routine; there's no
     * input parameter; nor is there in the SOAP version of this
     * query.  However, there /is/ a way to pass that information on
     * the command-line -- though that codepath is completely
     * untested.  I can see that it might be useful to provide this
     * information, but at present it's probably completely unused, so
     * the compromise for now is to hardwire the 16.
     */ 
    if(trackTable[k] < 16) {
      dudCount++; 
      if(!trackTable[k]) {
	nullCount++; 
      } 
    }
  }

  if(adb->header->length > 0) {
    if(munmap(trackTable, trackTableLength)) {
      return 1;
    }
  }

  status->numFiles = adb->header->numFiles;
  status->dim = adb->header->dim;
  status->length = adb->header->length;
  status->dudCount = dudCount;
  status->nullCount = nullCount;
  status->flags = adb->header->flags;
  status->data_region_size = adb->header->timesTableOffset - adb->header->dataOffset;

  return 0;
}