[Pkg-opt-media-team] Bug#615978: Additional information

[michael]

Hi Rogerio,

Sorry for nailing you down ! I suggested an overfow in the first place.

   Sharing experiences is cool, here are mine :

   We switched to BD-R several years ago, using growisofs to burn 
pregenerated pools of the correct size, burning them 'dao'.
   By request of a customer we been switching to multisession writting 
using dvd-r, also using growisofs.
   After switching to multisession we run into the error above.

   In the following testcase i use the same device (MATSHITA BD-MLT 
SW-5583 1.00) as in the previous tests, burning a folder with 5 blobs of 
1G each multiple times with different folder names.

   The informations for the blank media are in blank.info.

   Before taking any action, an growisofs -F /dev/sr2 has been called.

   The initial burn command line was :

     growisofs -Z /dev/sr2 -iso-level 4 -R -input-charset iso8859-2 
-output-charset iso8859-2 -overburn -V 0000 burn

   all following command lines read :

     growisofs -m /dev/sr2 -iso-level 4 -R -input-charset iso8859-2 
-output-charset iso8859-2 -overburn -V 0000 burn

   We using the patched growisofs added :

        fprintf (stderr," capacity  : %"LLD"\n", capacity);
        fprintf (stderr," outoff    : %"LLD"\n", outoff);
        fprintf (stderr," tracksize : %"LLD"\n", tracksize);

     at line 2397 (growisofs.c), before 'if (capacity && progress.final 
 > capacity)' as you implied,

   and growisofs_mmc.cpp, line 426, after 'ret = nwa + free_blocks;'

         fprintf (stderr," nwa       : %lli\n", nwa);
         fprintf (stderr," free blks : %lli\n", free_blocks);
         fprintf (stderr," result    : %lli\n", ret);

   after command returned, dvd+rw-mediainfo been executed.

   To prevent my keybord from damage, the qad script burn-one.sh is used.

   I hope you could extract the information you needed from the files 
attached.

   In any other case : ask, you're welcome.

   Tomorrow i will do same without parameter overburn, so if you have 
any suggestions ...

   regards,

     michael

p.s.:

   for the first session :

     -F :
	next_session=0
	capacity=50050629632
	next_session=0
	capacity=50050629632

     -Z :
  	nwa       : 0
	free blks : 24438784
	result    : 24438784
	Warning: Creating ISO-9660:1999 (version 2) filesystem.
	Warning: ISO-9660 filenames longer than 31 may cause buffer overflows 
in the OS.
   	 0.19% done, estimate finish Tue Mar  8 10:45:08 2011
   	 0.38% done, estimate finish Tue Mar  8 10:45:08 2011
	<snip>

   for the second session :

     -F :
  	nwa       : 2621632
  	free blks : 21686080
	result    : 24307712
	next_session=5369102336
	capacity=49782194176

     -Z :
  	nwa       : 2621632
	free blks : 21686080
	result    : 24307712
	Warning: Creating ISO-9660:1999 (version 2) filesystem.
	ISO-9660 image includes checksum signature for correct inode numbers.
	SUSP signatures version 1 found
	Rock Ridge signatures version 1 found
	Rock Ridge id 'RRIP_1991A'
	 50.06% done, estimate finish Tue Mar  8 11:03:54 2011
	 50.16% done, estimate finish Tue Mar  8 11:03:54 2011
	<snip>

   for last session :

  	nwa       : 23594944
  	free blks : 712768
  	result    : 24307712
	Warning: Creating ISO-9660:1999 (version 2) filesystem.
	Warning: ISO-9660 filenames longer than 31 may cause buffer overflows 
in the OS.
	SUSP signatures version 1 found
	Rock Ridge signatures version 1 found
	Rock Ridge id 'RRIP_1991A'
	:-( /dev/sr2: 712768 blocks are free, 2621626 to be written!
	:-! ignoring...
	 90.00% done, estimate finish Tue Mar  8 14:26:13 2011
	 90.02% done, estimate finish Tue Mar  8 14:26:14 2011
	<snip>
    	 92.77% done, estimate finish Tue Mar  8 14:29:49 2011
	:-[ WRITE at LBA=172e800h failed with SK=5h/INVALID ADDRESS FOR WRITE]: 
Invalid argument
	:-( write failed: Invalid argument
	/dev/sr2: flushing cache
	/dev/sr2: closing track

    medium mount's, lacking data from last session

p.p.s:

   By the way, are there any guesses about the error reported closing 
the first session :

    :-[ CLOSE SESSION failed with SK=5h/INVALID FIELD IN CDB]: 
Input/output error

   The net researches gave no clue about that, imho ignoring this error 
seems to be harmless, so maybe it should be suppressed.