Re: cdrecord and "get performance" scsi errors

["Thomas Schmitt" <scdbackup@gmx.net>]

Hi,

Bill Davidsen wrote:
> Is there some way [...] the "right" value for "i" can be determined?

If it is really true that Giulio's system+drive demand
a "Maximum Number of Descriptors" which is not larger
than the available number of descriptors, then the method
of dvd+rw-tools would be such a way:
- First execute the command with requesting at most 1
  descriptor (0 is suggested by MMC-5).
- Read the number of available descriptors from the reply.
- Repeat the command allowing exactly that number of descriptors.

The other two MMC commands of which Giulio reports problems
> "read track info" "read disk info"
are
  51h READ DISC INFORMATION
  52h READ TRACK INFORMATION
which both have a parameter "Allocation Length" with similar
purpose as parameter "Maximum Number of Descriptors" of 
our friend
  ACh GET PERFORMANCE

With 52h the reply size changed from 36 bytes in MMC-3 to
48 bytes in MMC-5.
The failing command of cdrecord allowed 48 bytes of reply.


This is the first time i see a reason for the careful
reply size preparations of growisofs. Up to now i wondered
why Andy does such an effort with replies where one can easily
make a sufficient estimation.

Well, i got my own riddle to explore. (Especially why telltoc
and cdrskin show a different number of commands although they
sit on the same same function in libburn.)


Have a nice day :)

Thomas