>>>>> "Niklas" == Niklas Hallqvist <niklas@appli.appli.se> writes:

Niklas> Hi, again.  I've started to get this message: gtsc_dma0:
Niklas> timeout #1 sometimes during startup.  I have recently changed
Niklas> my config so it'll configure my 2nd gtsc controller, but I
Niklas> don't know if it started before or after that change.  I
Niklas> haven't seen any bad behaviour apart from the message.  Before
Niklas> I look into it, does anyone know if this is normal when having
Niklas> two gtsc controllers configured.  If it had been gtsc_dma1 I'd
Niklas> not worried at all (hey that SCSIbus is empty and even
Niklas> unterminated) but it says gtsc_dma0, which is the controller
Niklas> where I keep all my data.

I looked into this this morning and found out why.  The message is
harmless.  It is caused by having two GVP controllers will result
in two watchdog timers running and *both* timer will check *both*
controllers status, thus the timeout counter will be increased *twice*
and the watermark will be reached easily.  There are two ways to fix
this:

A	Ensure only one watchdog is running whatever the count of
	gtsc devices there are.

B	Run one watchdog for each gtsc device, all of which checks
	just *one* controller.

Locally, I solved it using B, but I leave it up to Chris to say
what he likes most.  I'll happily provide patches for any of the two
solutions when you decide which we'll use, Chris.

Niklas

------------------------------------------------------------------------------