[Libcdio-devel] maybe i need more sleep, but isn't this a problem too?

[Robert William Fuller]

In mmc.c's mmc_read_cd(), shouldn't the CDIO_MMC_SET_READ_LENGTH24(...) 
be inside the loop?  Otherwise, reads exceeding MAX_CD_READ_BLOCKS will 
be incorrect?  (Also, shouldn't CDIO_MMC_SET_READ_LENGTH24 be using 
i_blocks2?)

excert from mmc.c's mmc_read_cd():

  CDIO_MMC_SET_READ_LENGTH24(cdb.field, i_blocks);


  if (b_sync)      cdb9 |= 128;
  if (b_user_data) cdb9 |=  16;
  if (b_edc_ecc)   cdb9 |=   8;
  cdb9 |= (header_codes & 3)         << 5;
  cdb9 |= (c2_error_information & 3) << 1;
  cdb.field[9]  = cdb9;
  cdb.field[10] = (subchannel_selection & 7);

  {
    unsigned int j = 0;
    int i_ret = DRIVER_OP_SUCCESS;
    const uint8_t i_cdb = mmc_get_cmd_len(cdb.field[0]);

    while (i_blocks > 0) {
      const unsigned i_blocks2 = (i_blocks > MAX_CD_READ_BLOCKS)
        ? MAX_CD_READ_BLOCKS : i_blocks;
      void *p_buf2 = ((char *)p_buf ) + (j * i_blocksize);

      CDIO_MMC_SET_READ_LBA (cdb.field, (i_lsn+j));

      i_ret = run_mmc_cmd (p_cdio->env, CD_READ_TIMEOUT_MS,
                           i_cdb, &cdb,
                           SCSI_MMC_DATA_READ,
                           i_blocksize * i_blocks2,
                           p_buf2);