Re: Issues with cdrsin and USB devices on RHEL5

[Bill Davidsen <davidsen@tmr.com>]

Thomas Schmitt wrote:
> Hi,
>
>   
> > I'm going to cautiously say that this file, intended for humans but can be
> > parsed, seems to handle anything up to a mix of IDE, "real" SCSI, and USB,
> >     
>
> Yep. For me it yields plausible results with
> 4 drives at USB, IDE, and SATA.
>   drive name:             sr1     hdc     hda     sr0
>
>
>   
> > and tells you what the kernel thinks of them. That may be more important
> > than the truth, if a device says it can support feature X, but the kernel
> > doesn't agree, I would not expect that feature to work reliably.
> >     
>
> If the CDROM driver believes it then it does not
> necessarily have to be true.
> Example:
>   Can read multisession:  1       1       1       1
> is true with hda only for CD but not for DVD media.
> This DVD-ROM drive is too stupid to recognize multiple
> sessions on DVD-R or DVD+R. (One can scan for ISO 9660
> headers, though, and successfully mount the sessions
> via option sbsector= .) 
>
>   
Wow, is that actually useful for anything? Amazing!

I guess the point is that the drive has support for multisession and 
advertises same. If this value was zero, I doubt that any option would 
help. Firmware bugs are not known to the kernel, unfortunately. I would 
treat this as and AND of capabilities, if this file says the capability 
is missing it probably really is. If present, the device supports the 
feature or lies convincingly. ;-)
> So the feature detection is a direct competitor
> of libburn's own view of things. To take it into
> respect would demand to establish conflict handling
> in case of contradictions.
>
> Nevertheless, i will consider to take the "drive name:"
> line as a hint of how the drive list should look like.
> In case of deviations, one could issue a warning.
>
>   
Yes, that certainly should be useful in eliminating the non-optical media.
>   
> > Thought it might be useful to tell you what the kernel
> > would let you see and use.
> >     
>
> A totally overhauled drive detection strategy will have
> to collect info from such sources and try to verify
> it by own drive inquiries.
>
> So any hint about CD drive info sources is welcome.
>
> For now i seem to have put a patch on the problems.
> After the upcoming release i will reconsider that
> topic and make some larger changes in the Linux adapter
> of libburn.
>   

--
Bill Davidsen <davidsen@tmr.com>
 "Woe unto the statesman who makes war without a reason that will still
 be valid when the war is over..." Otto von Bismark