So this is some hardware issue; as long as nobody else cares about this, I won't 
do anything about it.

Michael Prokop <mika@grml.org> added the comment:
> * Christian Hofstaedtler wrote in grml's BTS on 20110320 / 13:14:
>> We should at least document what the underlying issue is, and if there was a bug
>> in earlier grml2usb versions, than that too.
>
> grml2usb checks for /sys/class/block/sdX/removable and if there's
> "1" inside it's detected as removable device, otherwise not.
> Problem: some usb keys (wrongly) claim to be non-removable devices
> (the Grml key suffers from this for example as well).

Yeah, but there seems to be an additional problem somewhere:
Had a Grml key and some other USB key.
On 2010.04 with grml2usb from grml2usb.tgz (0.9.28), both keys were
thought to be non-removable.
On 2010.12 with the shipping grml2usb only the Grml key suffered from
this problem, the other one was fine.
My guess would be that the kernel also changed under us.

> Could we safely consider usb drives to be *always* removable devices?

Possibly. Can we reliably detect usb drives? (driver == usb-storage?)

  -ch

* Christian Hofstaedtler wrote in grml's BTS on 20110320 / 13:14:

> grml2usb sometimes complains that the target isn't a removable device, when this 
> is clearly not the case.

Yes.

> We should at least document what the underlying issue is, and if there was a bug 
> in earlier grml2usb versions, than that too.

grml2usb checks for /sys/class/block/sdX/removable and if there's
"1" inside it's detected as removable device, otherwise not.
Problem: some usb keys (wrongly) claim to be non-removable devices
(the Grml key suffers from this for example as well).

Could we safely consider usb drives to be *always* removable devices?

regards,
-mika-

grml2usb sometimes complains that the target isn't a removable device, when this 
is clearly not the case.
We should at least document what the underlying issue is, and if there was a bug 
in earlier grml2usb versions, than that too.