Reattach/redetect allways connected umass device - is it possible ?

[Bernd Walter]

On Mon, Mar 28, 2005 at 06:06:10PM +0400, Vladimir Grebenschikov wrote:
> ? ??, 28/03/2005 ? 15:28 +0200, Bernd Walter ?????:
> > On Mon, Mar 28, 2005 at 05:17:24PM +0400, Vladimir Grebenschikov wrote:
> > > ? ??, 28/03/2005 ? 15:04 +0200, Bernd Walter ?????:
> > > > On Mon, Mar 28, 2005 at 04:58:31PM +0400, Vladimir Grebenschikov wrote:
> > > > > ? ??, 28/03/2005 ? 14:13 +0200, Poul-Henning Kamp ?????:
> > > > > > In message <20050328114633.GZ14532 at cicely12.cicely.de>, Bernd Walter writes:
> > > > > > 
> > > > > > >> camcontrol detach da0; camcontrol rescan all 
> > > > > > >> helps, but, it should be much better if it will be issued automatically.
> > > > > > >
> > > > > > >Yes - GEOM seems to ignore media change signals from drives.
> > > > > > >I've added PHK to the recipient list - maybe he has an idea about this
> > > > > > >problem.
> > > > > > 
> > > > > > No, GEOM doesn't ignore any such thing, because as far as I know
> > > > > > GEOM doesn't get any such thing to ignore in the first place.
> > > > > 
> > > > > So, let's imagine following situation:
> > > > > 
> > > > > We get SCSI BUS with removable da device.
> > > > > device detected as da0 and not mounted.
> > > > > Device disconnected from SCSI bus.
> > > > > And finally, another device with different geometry connected with same
> > > > > SCIS ID.
> > > > 
> > > > This ist not a *media* exchange - this is a *device* and in
> > > > this case even a scbus exchange.
> > > 
> > > Ok, so my case is media exchange, not device exchange.
> > 
> > No - it is a device exchange.
> > GEOM only has problems with media exchange.
> > A new device is no problem for GEOM.
> > Your problem is very likely not GEOM related.
> 
> I guess, may bad example messed up discussion. sorry.
> 
> >From begin:
>   I have USB-connected card-reader (Sony Memory Stick).
>   It connected as umass0 (ehci), then as da0.
>   It can't be unplugged. (on-board notebook device)

OK - then I got you wrong here - so you are really exchanging just
the media.
Yes - the thread is very mixed - and I don't imagine a memory stick,
which describes the physical design IMO, to be an onboard device.

> Problem:
>   When I change card (it was not mounted) da0 device does not detect
>   media change and following mount fails.

Ack.

> Workarounds (I know three at this moment):
>   - kldunload umass; kldload umass - the ugliest, will affect other
> umass devices.
>   - cat /dev/null > /dev/da0  before mount - GEOM hacking 
>     (also I do not like do cat into device with data)

I think this is currently the official way.
It just does what is required to trigger a GEOM rescan.
Everthing else removes the complete device, which is more than
required.

>   - camcontrol detach da0; camcontrol rescan all - requires CAM patch
> 
> So, I am asking about solution, how it is supposed to work, 
> I guess, like with CD, there is should not be special exercises after
> card change.

If I got Poul right then even for the CD case it just works
because of some evilness.

> Probably my card-read is too stupid to send any media change
> notifications, how to test it ?

You should see SCSI messages from the kernel for the first access
after exchanging the media.
But from what I know GEOM doesn't know how to handle exchange
notifications from SCSI devices.
What we IMHO really need is to make the da driver poll the device
for media exchange notifications - and hoope that the drive is
capable to do.
I don't have an idea for drives that don't support notifications.

> Another, probably related question, is there way to detect card
> insertion/removal ?

You can explizitly ask a SCSI device for media and readyness state.

-- 
B.Walter                   BWCT                http://www.bwct.de
bernd at bwct.de                                  info at bwct.de