Re: [Libcdio-devel] Read Sub-channel changes

[Rocky Bernstein]

>
> Enumeration, Macros ? Are there any proposals for symbol names ?

  I need to express this outcome
>            0 = no problem


Yes, use an enumeration. Look at include/cdio/cdtext.h for an example. And
just preface each value in the enumeration with something that seems right,
for example SENSE_KEY_ASC_ASCQ_... which is a little long, so feel free to
come up with feels right to you.

As for how close to follow SPC-3 and your experience etc, you decide. You
are overthinking this. Instead of trying to get it perfect the first time,
make some decisions and plan on adjusting things as usage and experience
suggests.


I do not propose to remove it from the API.
> It's just that  mmc_last_cmd_key_asc_ascq()  does its job without
> the caller needing to know about SPC and sense data formats.


However unlikely, it is possible that someone in the world has an idea
about or knows a better way to make sense of the sense keys. Or perhaps
that is already been done somewhere and we just don't know about it. So by
keeping around the more raw information, code that might have been written
to use it still works and experimenters can try things using the more
one-to-one representation and work outside of the libcdio code base using
libcdio as a black box.

On Tue, Jun 7, 2016 at 10:29 AM, Thomas Schmitt <address@hidden> wrote:

> Hi,
>
> i wrote:
> > > improved replacement of
> > > mmc_last_cmd_sense() and cdio_mmc_request_sense_t,
>
> Rocky Bernstein wrote:
> > For compatibility let's think of keeping both around at least for a
> release
>
> I do not propose to remove it from the API.
> It's just that  mmc_last_cmd_key_asc_ascq()  does its job without
> the caller needing to know about SPC and sense data formats.
>
>
> > > prototype of mmc_sense_code_to_text()
> >  put it in  include/cdio/mmc_util.h
>
> Now i wonder why i did not find this file when looking for a place.
> (Probably because i looked in lib/driver/mmc rather than in include/cdio.)
>
> Ok. The prototype is in include/cdio/mmc_util.h .
>
> -------------------------------------------------------------------------
>
> About the evaluation of sense codes:
>
> I am testing a function now. Some questions arise.
>
> - How to represent the result in the architecture and style of libcdio ?
>   Enumeration, Macros ? Are there any proposals for symbol names ?
>
>   I need to express this outcome
>
>            0 = no problem
>            1 = problem which needs no direct reaction
>                (but it may announce particular future failures if not
>                 proper action is taken)
>            2 = desired action failed, retry may bring success
>                (timeout or abort due to repeated failure are responsibility
>                 of the caller)
>            3 = desired action failed, retry seems futile
>            4 = relation to drive is broken
>
>   and this instruction which is given by the caller
>
>            0 = no report by evaluator
>            1 = report as debug
>            2 = report as warning
>            3 = report as error
>
>
> - How much shall my user experience with the Linux kernel influence
>   my reading of SCSI specs ?
>
>   SPC-3 says that key 0xB invites for retry.
>   My (mostly hearsay) experience is that key 0xB indicates a broken
>   relation between kernel and drive.
>
>   Shall the evaluator really follow SPC-3 and advise retry ?
>   (I fear that the next SCSI command might not return.)
>
> -------------------------------------------------------------------------
>
> Have a nice day :)
>
> Thomas
>
>
>