Re: sata_mv (NOT!)

[Mark Lord <liml@rtr.ca>]

Mark Lord wrote:
> Jeff Garzik wrote:
>> Mark Lord wrote:
>>>
>>> Yeah, it looks to me as if libata still has DVD burner issues.
>>> I used my SATA DVDRW drive to burn a Suse DVD the other day,
>>> and the same drive refused to read the burnt disc without first 
>>> rebooting.
>>>
>>> linux-2.6.25.6
>>>
>>> Disc was fine, though.  Single-layer.
>>
>> Do you get error messages as well?  Of what kind?
> ..
> 
> There have been so many broken things in so many parts of the kernel of 
> late,
> that I'm just way too far behind in diagnosing/fixing them.   Or 
> reporting them,
> since many developers have stated they would rather not see a bug report
> without a bisect.
> 
> But here's all that was in the logs for this particular bug.
> These messages appeared when trying to read the freshly burned DVD,
> both before and after eject/reinsert.  A reboot cured it:
> 
> [ 8209.473181] attempt to access beyond end of device
> [ 8209.473189] sr0: rw=0, want=68, limit=4
> [ 8209.487646] attempt to access beyond end of device
> [ 8209.487655] sr0: rw=0, want=5480256, limit=4
..
> [ 8209.514467] attempt to access beyond end of device
> [ 8209.514467] sr0: rw=0, want=68, limit=4
> [ 8209.514467] isofs_fill_super: bread failed, dev=sr0, iso_blknum=16, block=16
..

To hazard a guess at the problem here, I'd say that perhaps the media change
went undetected, and so sr_media_change() never got called.

I'm no longer even sure what would trigger that call.

Cheers