Re: Double write speed for BD-R without defect management

[Thomas Schmitt]

Hi,

carsten.kunze@arcor.de wrote:
> $ dvd+rw-format -ssa=268435456 /dev/sr0
> ...
> - illegal command-line option for this media.

Hm. I see in my experimental copy that i added profile 0x41 (BD-R) to the
test for "ssa"-suitable media before that message:

===========================================================================
@@ -699,9 +704,11 @@ int main (int argc, char *argv[])
        force = 0;
     }

+/* ts B70205 : added 0x41 as applicable profile */
+
     if (((mmc_profile == 0x1A || mmc_profile == 0x2A) && blank)
        || (mmc_profile != 0x1A && compat)
-       || (mmc_profile != 0x12 && mmc_profile != 0x43 && ssa) )
+       || (mmc_profile != 0x12 && mmc_profile != 0x43 && mmc_profile != 0x41 
&& ssa) )
     {  fprintf (stderr,"- illegal command-line option for this media.\n");
        goto offer_options;
     }
===========================================================================

I don't find matching mails which explain why i had reason to remove this
obstacle.
It is obviously unintended, because around line 420 there is a lengthy
handling of ssa sizes for BD-R:

    else if (mmc_profile==0x41)         // BD-R
    { unsigned int   max,min,cap;

        switch (ssa)
        { ...various cases...


> - BD-R can be pre-formatted only once

This message seems to be emitted on any refusal without refering to the
current state of the medium.


> xorriso -outdev /dev/sr0 -speed 2 -as cdrecord -v image.udf
> does write with effective 4.1x speed?

The write speed is finally a decision of the drive. The burn program can
make a proposal by SCSI command SET STREAMING.
xorriso lets libburn send what the user desires.

I did with a BD-RE

  xorriso-1.4.6 -scsi_log on -outdev /dev/sr4 -speed 4 \
                -as cdrecord -v debian-10.6.0-amd64-netinst.iso 2>&1 | less

and see in the (very lengthy) output

  SET STREAMING
  b6 00 00 00 00 00 00 00 00 00 1c 00
  To drive: 28b
  00 00 00 00 00 00 00 00 00 b8 73 ff 00 01 5f 39 00 00 03 e8
  00 00 46 3f 00 00 03 e8

The decisive bytes are the last 8:

  Write Size = 0x0000463f = 17983
  Write Time = 0x000003e8 = 1000

The MMC-5 specs say:

  The Write Size field shall be set to the number of kilobytes to be
  written per Write Time.
  The Write Time field shall indicate the amount of time, in milliseconds,
  over that the Write Size is expected to be written.

MMC obeys SI rules. So a "kilobyte" is 1000 bytes and thus the SCSI command
asks for 17,983,000 bytes per second.
BD 1x speed is 4,496,000 bytes per second. Division yields 3.999778.
So my speed wish was computed quite exactly and sent to the drive.

Just the drive did not obey and wrote at 2.0x speed, as expectable with
a BD-RE.

What report do you get with the formatted medium in the drive from this
xorriso run:

  xorriso -outdev /dev/sr0 -list_speeds

With an unformatted, partly written BD-R i get:
  ...
  Write speed  :  53952k , 12.0xB
  Write speed  :  44960k , 10.0xB
  Write speed  :  35968k ,  8.0xB
  Write speed  :  26976k ,  6.0xB
  Write speed  :  17984k ,  4.0xB
  Write speed L:  17984k ,  4.0xB
  Write speed H:  53952k , 12.0xB

So i will probably not get fulfilled a wish for less than 4x.
I expect that it will give me the lowest speed from its repertoire.

(Hopefully i will remember tomorrow to try -speed 2 when the next session
is to be written to the medium.)


> My xorriso version is
> xorriso 1.4.6 : RockRidge filesystem manipulator, libburnia project.
> ...
> Version timestamp :  2016.09.16.133001

This is not the freshest version. But such elementary things as speed
setting should not have changed in the last four years.


Have a nice day :)

Thomas