Re: [Libcdio-devel] [PATCH] Try ejecting using eject command on Linux

[Robert Kausch]

Turns out it's all much simpler...

We are opening devices in O_RDONLY mode by default, but to run the 
PREVENT_ALLOW_MEDIUM_REMOVAL MMC command on Linux, we need to open it 
with O_RDWR. I changed eject_media_linux to make sure the device is 
opened with O_RDWR and everything works beautifully now.

The patch is quite simple, so I already pushed it to git. I will check 
if other OS need similar adjustments in the next few days.

Robert

Am 12.06.2015 um 19:15 schrieb Thomas Schmitt:
> Hi,
>
> > Maybe CDROM's now have a "suspended" state that you need get out of?
> It could be about "standby power condition" (SPC-3 5.9).
> Table 39 says:
> "While in the standby power condition (see 3.1.107):
>    a) A device server is not capable of processing media access commands; [...]
> "
>
> But on the other hand, START/STOP UNIT is supposed to override
> resp. change the power condition. Especially the value 0 of bits
> 4 to 7 of byte 4 is promised to "process the START and LOEJ bits"
> which are bits 0 resp. 1 of byte 4. (LOEJ = Load/Eject)
> SBC-2 table 49.
>
> I have this reminder in libburn/sbc.c:
>
> /* START STOP UNIT as of SBC-1 and SBC-2
>     0:  Opcode 0x1B
>     1:  bit0= Immed
>         bit1-7= reserved
>     2:  reserved
>     3:  reserved
>     4:  bit0= Start (else Stop unit)
>         bit1= Load/Eject (according to Start resp. Stop)
>         bit2-3= reserved
>         bit4-7= Power Condition
>                 0= Start Valid: process Start and Load/Eject bits
>                 1= assume Active state
>                 2= assume Idle state
>                 3= assume Standby state
>                (5= SBC-1 only: assume Sleep state)
>                 7= transfer control of power conditions to logical unit
>                10= force idle condition timer to 0
>                11= force standby condition timer to 0
>                All others are reserved.
>     5:  Control (set to 0)
> */
>
> The sequence which libburn uses with
>    xorriso -scsi_log on -outdev /dev/sr0 -eject all
> is inconclusive in this aspect. It has a START UNIT command which
> should bonk the drive out of its sleep, before anything else
> happens.
>
>    ...
>
>    START/STOP UNIT
>    1b 00 00 00 01 00
>        15070 us     [ 1096565 ]
>
>    ... some drive inspection ...
>
>    START/STOP UNIT
>    1b 01 00 00 00 00
>      1186603 us     [ 2406124 ]
>
>    TEST UNIT READY
>    00 00 00 00 00 00
>    +++ sense data = 71 00 02 00 00 00 00 0A 00 00 00 00 04 07 00 80 FF FF
>    +++ key=2  asc=04h  ascq=07h
>        23600 us     [ 2930067 ]
>
>    TEST UNIT READY
>    00 00 00 00 00 00
>    +++ sense data = 71 00 02 00 00 00 00 0A 00 00 00 00 04 07 00 80 FF FF
>    +++ key=2  asc=04h  ascq=07h
>        53081 us     [ 3483470 ]
>
>    TEST UNIT READY
>    00 00 00 00 00 00
>          306 us     [ 3983994 ]
>
>    PREVENT/ALLOW MEDIA REMOVAL
>    1e 00 00 00 00 00
>          242 us     [ 3984516 ]
>
>    START/STOP UNIT
>    1b 00 00 00 02 00
>      1547037 us     [ 5531605 ]
>
> I.e. start unit with tray action (= load),
> stop unit without tray action,
> unlock tray,
> stop unit with tray action (= eject).
> After this command the tray is out. (Needs 1.5 seconds. :))
>
> The final command from libburn confirms this:
>
>    START/STOP UNIT
>    1b 01 00 00 00 00
>    +++ sense data = 70 00 02 00 00 00 00 0A 00 00 00 00 3A 02 00 00 00 00
>    +++ key=2  asc=3Ah  ascq=02h
>          946 us     [ 5532752 ]
>
> Sense codes seen:
>    2 04 07 LOGICAL UNIT NOT READY, OPERATION IN PROGRESS
>    2 3A 02 MEDIUM NOT PRESENT  TRAY OPEN
>
>
> >   ftp://www.t10.org/t10/document.05/05-344r0.pdf
> Oops. SBC-3. I thought they want money since a few years.
> Table 49 is in there.
> Maybe SPC-3 (or SPC-4) is hidden there too.
>
>
> Have a nice day :)
>
> Thomas