[Stgt-devel] I have found a small problem in the MMC code ...

[Richard Sharpe]

Hi,

I have found what I think is a small problem in the MMC code, although
I am not sure.

I am trying to use cdrecord to write to a DVD Jukebox I have set up
following Ronnie's example and my code changes.

When the client asks for the TOC, it seems to give us too small an
allocation length (four bytes) and of course, we give it back only
what is asked for ... it looks like maybe cdrecord is doing this,
because the Linux kernel seems to offer a buffer of 12 (at least that
is what the code says).

In any event, this causes the first write that CD record requests to
return a CHECK_CONDITION. cdrecord then recovers and the writing of my
8GB DVD :-) completes ...

I will have to check the specs carefully, but I wonder if we should be
returning CHECK_CONDITION as the response to the request for the
request to read the TOC/PMA/ATIP or is the client supposed to check
that there was more data than the buffer could hold and then reissue
the command?