APMD-List:
Archives

  From: Craig Markwardt <craigm@pcasun3.gsfc.nasa.gov>
  To  : <apmd-list@worldvisions.ca>
  Date: Fri, 18 Jun 1999 15:03:59 -0400 (EDT)

Re: apm daemon, suspend/resume

[ Discussion with PCMCIA maintainer David Hinds forwarded with permission. ]

Craig Markwardt writes:
 > 
 > 
 > David Hinds writes:
 >  > You mean "cardctl suspend" and "cardctl resume", right?
 > 
 > Yes.
 > 
 >  > > I am also aware that the PCMCIA kernel modules directly receive APM
 >  > > events and respond to them.  However, cards are hard suspended,
 >  > > right(?)
 >  > 
 >  > What do you mean by "hard suspended"?
 >  > 
 >  > When a PCMCIA client gets a suspend signal, it is supposed to cease IO
 >  > to the card, and the core modules then power down the socket.  When a
 >  > resume is received, the socket is powered up and the client performs a
 >  > full reset of the device.
 > 
 >  > For PCMCIA, the cardmgr daemon actually receives the suspend event and
 >  > does the equivalent of a "cardctl suspend".  So I'm surprised that it
 >  > makes any difference if you have apmd do the same thing: are you sure
 >  > that this has any effect?
 > 
 > Are you saying that "cardctl suspend" and an APM suspend event handled
 > by the PCMCIA core modules will result in the same thing?  I don't
 > think that's what people have been experiencing.  Personally, I don't
 > have problems, but every once in awhile I've seen messages from people
 > who have to be sure to "suspend" (or maybe "eject"?) their card before
 > going into suspend mode.
 > 
 > The question is whether "cardctl suspend" should be called by apmd by
 > default.  It's in a shell script, so people can configure it.  I've
 > been arguing against this as a default behavior, since I think it will
 > probably fry more people that it will help.  My personal feeling is
 > that the script should do nothing as a default, so as to provide a
 > minimum of interference for a new user.
 > 
 > Craig
 > 
 > P.S.  May I forward this correspondence to the apmd mailing list?
 > Thanks.
 >