On Sun, Sep 26, 2010 at 04:04:45PM +0000, Miod Vallat wrote:
quoted text
> > New diff for the simple pms/pmsi merge.
> 
> Whey did you get rid of the poll- vs interrupt-driven operation in
> pms_setintellimode()? It is necessary for proper suspend/resume
> operation.
> 
> Miod

In original pms_intelli.c function pmsi_setintellimode() 
called with the poll = 1 only in pmsiprobe() and pmsiattach().
In pmsi_change_state() poll = 0.

In new pms/pmsi function pms_setintellimode() called only in
pms_change_state(). That's why.

--
Alexandr Shadchin