Hello.Justin Madru wrote:After referring to the SCSI command codes "cdb 0x1e" means ALLOW MEDIUM REMOVAL command -- which
could hardly be addressed to an usual hard disk. So, it looks like we had a case of the confused bug report which
has a lot of info on the hard disk while errors were most probably happening with a CD/DVD drive.
Yes, I originally thought it was my hard disk because the kernel logs showed ata2.
But, Tejun Heo figured out it was my DVD drive (ATAPI) that was on the ata2 link.
(see http://marc.info/?l=linux-kernel&m=122993014109646&w=2)
I tried to bisect it, but around .28-rc1 I began to get numerous compile errors, so couldn't continue.
I also tried patches that Tejun sent me, but non of them worked, it just slightly change the error message.
So, yes this is a regression that was introduced in the .28 merge window, and I still think that bug #12609 is a duplicate of my bug.
If 12609 is truly a post-2.6.28 regression and 12263 is post-2.6.27 regresssion, this just cannot be.
I don't see this bug on tip/master and this is the diff of origin and tip at the time I tested.
$ git diff origin/master..tip/master drivers/ata/
What tree is that?