Re: [Libcdio-devel] Burned CD-RW and DVD+RW via libcdio MMC interface

[Rocky Bernstein]

You are *not* annoying. Please keep up the good work! We are interested in
the bug fix. I have been busy work that I can't stop and devote the
attention that is due this.

I hope sometime this weekend I hope I will be able to respond more in full.

Thanks for your patience.

On Sat, Dec 19, 2009 at 10:43 AM, Thomas Schmitt <address@hidden> wrote:

> Hi,
>
> at the risk to become annoying i want to report
> that i managed to blank a CD-RW and to write a
> session of Mode 1 sectors in TAO mode via
> libcdio's mmc_run_cmd(). I also wrote to a
> DVD+RW.
> This is implemented as experimental libburn
> system adapter os-libcdio.h and sg-libcdio.c.
> (libburn SVN revision 2959)
>
> Both images verified against their recorded MD5
> checksums afterwards.
>
> But:
>
> - I had to make changes in libcdio:
>
>  - Fix the reported bug about data transfer
>    direction in
>      run_mmc_cmd_linux()
>
>  - Introduce as third transfer direction
>    SCSI_MMC_DATA_NONE
>
>  - Open in cdio_open_am_linux() the drive by
>      cdio_generic_init(_data, O_RDWR|O_NONBLOCK)
>    rather than O_RDONLY.
>
> - An attempt to write a CD SAO session failed.
>  I did not yet find out why.
>
> - mmc_run_cmd() does not give me the SCSI sense
>  reply. libburn interprets Key,Asc,Ascq
>  to learn what went wrong. Not only in case of
>  severe errors, but also when performing
>  try-and-error gestures.
>  Without Key,Asc,Ascq the new adapter
>    libburn/sg-libcdio.c
>  has to issue a dummy reply in case of a failed
>  MMC command.
>  (Currently this keeps libburn from ending after
>   eject because it waits for the
>   Media-Not-Present reply and always gets the
>   dummy Not-Ready reply.)
>
> ------------------------------------------------
>
> Would you be interested in my bug fix ?
>
> Are there experts who could implement direction
> SCSI_MMC_DATA_NONE and write permission in the
> various MMC drivers and maybe test libburn on
> those drivers ?
> (Both changes are not only necessary for libburn
> but for any attempt to use libcdio for writing,
> i believe.)
>
> Would "access mode" of cdio_open_am() be a
> way for controlling O_RDONLY versus O_RDWR ?
> (E.g. a new mode "MMC_RDWR")
>
>
> Have a nice day :)
>
> Thomas
>
>
>
>