Guest software dealing with CD/DVD drives use this command (GPCMD_READ_DISC_INFO in /usr/include/linux/cdrom.h).  Not all software is prepared for this command missing.  It is known to have tripped up anaconda (bug 496298) and udev (bug 647794).  I'm not aware of any problems caused by the command missing at this time.  But it would be nice to have.

However, we'd need exception or blocker to fix this for 6.2, and that can't be justified as long as this is just "nice to have".  Deferring to 6.3.

MMC says that the command only needs to be implemented if any of the following features are reported in GET CONFIGURATION:

Feature Number Feature Name                       Command Requirement
001Dh          Multi-Read                         Mandatory
0021h          Incremental Streaming Writable     Mandatory
0026h          Restricted Overwrite               Mandatory
0027h          CD-RW CAV Write                    Mandatory
0029h          Enhanced Defect Reporting          Mandatory
002Ah          DVD+RW                             Mandatory
002Bh          DVD+R                              Mandatory
002Ch          Rigid Restricted Overwrite         Mandatory
002Dh          CD Track at once                   Mandatory
002Eh          CD Mastering                       Mandatory
002Fh          DVD-R/-RW Write                    Mandatory
0038h          BD-R POW                           Mandatory
003Bh          DVD+R DL                           Mandatory

(MMC6, 6.21 READ DISC INFORMATION command, table 362)

None of these is required for CD-ROM and DVD-ROM profiles, so the spec grants us the possibility to WONTFIX this.  Markus, what do you think?

Upstream commit 55042b95 (ATAPI) and commit 1a4f0c3a (SCSI).  I haven't investigated how invasive a backport would be.

Original rationale for the RFE, from comment#1:

Guest software dealing with CD/DVD drives use this command (GPCMD_READ_DISC_INFO in /usr/include/linux/cdrom.h).  Not all software is prepared for this command missing.  It is known to have tripped up anaconda (bug 496298) and udev (bug 647794).  I'm not aware of any problems caused by the command missing at this time.  But it would be nice to have.

End quote.  

No additional problems have been reported.  Considering how late we are in the RHEL-6 cycle, CLOSED/DEFERRED should be seriously considered.

Backport is uninvasive, but this also means that if problems are actually reported we can do it fast.  A wholesale backport of upstream SCSI changes is unlikely to happen again in RHEL-6.